Skip to main content

README

Coding Agent

For every coding task: which agent, model and reasoning effort to use with your subscription, how to work with it and when to switch — to get the best result before your quota runs out.


What is a coding agent?

An AI that works inside your project: it reads the files, edits the code, runs commands and tests, and iterates until the task is done, with you reviewing. The same agent is often reachable in many ways — a terminal (CLI), a desktop app, an IDE extension, a web service, remotely from a phone or another browser, inside GitHub and CI/CD (pull request reviews, agents assigned to issues, GitHub Actions), from team tools (Slack, Teams, Linear, Jira…) or through an SDK in your own scripts: Codex, for example, runs in Codex CLI, the ChatGPT desktop app, the Codex IDE extension, Codex cloud, Codex Remote, Codex code review, Slack and the Codex SDK. Below, the most used agents, most used first; this page covers them as long as they are used through a subscription.

AgentUsage at workCLI (terminal)Desktop appIDE extensionWeb / chatRemote & mobileGitHub & CI/CDTeam toolsSDK & scripts
Claude CodeAnthropic39%npm 12.5M/wkClaude Code CLIClaude Code DesktopClaude Code for VS CodeClaude Code on the webRemote Control, Claude app (iOS, Android)Claude Code GitHub Actions, GitLab CI/CDClaude Code in SlackClaude Agent SDK, claude -p
GitHub CopilotGitHub21%npm 1.6M/wkGitHub Copilot CLIGitHub Copilot extensionCopilot Chat (github.com/copilot)GitHub MobileCopilot code review, Copilot cloud agentSlack, Microsoft Teams, Jira, Linear, Azure BoardsGitHub Copilot SDK, copilot -p
CodexOpenAI16%npm 23.1M/wkCodex CLIChatGPT desktop appCodex IDE extensionCodex cloudCodex Remote (ChatGPT mobile app)Codex code review (GitHub, GitLab preview), Codex GitHub ActionSlack, LinearCodex SDK, codex exec
CursorCursor12%Cursor CLICursor (editor)JetBrains (ACP)Cloud AgentsCursor for iOSBugbotSlack, LinearCursor SDK, headless CLI
JunieJetBrains9%with JetBrains AIJunie CLIJunie in JetBrains IDEsGitHub, GitLab
OpenCodeAnomaly7%npm 3M/wkOpenCode TerminalOpenCode DesktopOpenCode Extensionsopencode web (network access)GitHub, GitLabOpenCode SDK, opencode run
AntigravityGoogle6%Antigravity CLIAntigravity IDE, Antigravity 2.0Antigravity IDE ExtensionsRemote ControlAntigravity SDK
Qwen CodeAlibaba (Qwen)npm 68.8K/wkQwen Code CLIQwen Code DesktopQwen Code CompanionChannels: Telegram, WeChat, QQQwen Code GitHub ActionChannels: DingTalk, WeCom, FeishuQwen Code SDK, headless mode
Grok BuildxAInpm 60.6K/wkGrok Build CLIvia ACPHeadless mode
Kimi CodeMoonshot AInpm 47.8K/wkKimi Code CLIKimi Code DesktopKimi Code for VS CodeKimi Agent SDK
ZCode / Claude CodeZ.ai—Claude CodeZCodeClaude Code IDE PluginRemote Control, WeChat botFeishu bot
  • heavily used on that surface: at least 1 M CLI or SDK downloads a week, 1 M IDE-extension installs, 10,000 repositories using the GitHub action or 1 M pull requests reviewed a month; desktop, web, remote and team tools publish no figures, so only the editor the survey measures (Cursor, Antigravity) gets one there. Point at a green ✓ to see its figure.
  • available on that surface.

Usage: share of professional developers using the agent at work, May–July 2026 (JetBrains Developer Ecosystem Survey 2026, 15,000+ developers). npm: weekly downloads of the agent's CLI, October 4 – 10, 2026. Downloads include CI runs and automatic updates, so they only order the agents that the survey does not report. Qoder and MiMo Code are left out: fewer than 12,000 weekly downloads each. The gray names are the vendors' own names for each surface; the web column counts hosted services only, not local browser interfaces.

Sources for usage and surface names (73)

AI coding agent adoption 2026 · npm downloads API: coding agent CLIs, week of 2026-10-04 · VS Code Marketplace installs · JetBrains Marketplace downloads · GitHub Action dependents · Platforms and integrations · Use Claude Code Desktop · Claude Code for VS Code · Claude Code GitHub Actions · Use Claude Code in the cloud · npm downloads API: coding agent SDKs, week of 2026-10-04 · Continue local sessions from any device with Remote Control · Claude Code on mobile · Claude Code in Slack · Agent SDK overview · Run Claude Code programmatically · 60 million Copilot code reviews and counting · About GitHub Copilot CLI · Installing the GitHub Copilot extension in your environment · About GitHub Copilot code review · About GitHub Copilot cloud agent · Asking GitHub Copilot questions in GitHub · Asking GitHub Copilot questions in GitHub Mobile · Copilot SDK · A Practical Approach to Verifying Code at Scale · Codex glossary · Codex changelog: Codex joins the ChatGPT desktop app · Codex CLI · Codex IDE extension · Review GitHub pull requests with Codex · Codex cloud · Codex Remote · Use Codex in Slack · Use Codex in Linear · Codex SDK · Non-interactive mode · Cursor CLI · Cursor in JetBrains IDEs · Bugbot · Building a better Bugbot · Cloud Agents · Cursor for iOS · Cursor in Slack · Cursor in Linear · Cursor SDK · Using Headless CLI · Junie, the coding agent by JetBrains · Download OpenCode · OpenCode GitHub · OpenCode web · OpenCode SDK · OpenCode CLI · Google Antigravity docs: choose your surface · Antigravity CLI installation · Antigravity IDE Extensions · Antigravity Remote Control · Qwen Code · Qwen Code in VS Code · Qwen Code Desktop v0.24.6 · qwen-code-action · Qwen Code Channels · Headless Mode · Qwen Code TypeScript SDK · Grok Build · Grok Build headless and ACP · Kimi Code overview · Kimi Code Desktop changelog · Kimi Agent SDK · GLM Coding Plan supported tools · Install ZCode · Claude Code IDE Plugin · ZCode Remote Control · ZCode Bot Channel


Current recommendations

as of 2026-11-11

Data checked on November 11, 2026 · 277 sources

guide/what-to-use

What to use, when

Pick what you are about to do: you get the provider, model, reasoning effort and subscription to use, how long to stay on it and what to switch to next, how many agent-hours your quota gives — with the evidence behind each choice.

What do you want to do?

In the order of the work: plan first, then code and debug, then the bigger and more specialized jobs.

Advanced planning and architecture

Writing a development plan, a spec, a migration strategy, an architecture review.

  • Write a development plan before a feature
  • Design a migration (framework, database, API version)
  • Architecture or security review of a codebase
  • Turn a vague brief into a spec and task list

No benchmark measures the quality of a plan or a spec. The pick rests on general reasoning (AA Intelligence Index) and on LibraryDesignBench, which grades how well a model designs a library's API. Security reviews are a tie on CWE-bench v2 (250 tasks: Opus 5.5 67.1%, Gemini 4 Argon 65.0%, GPT-6 Astra 64.2%, ±5 points; Fable 5.2 and Gemini 4 Carbon not ranked yet).

Tied choices — pick the one your subscription gives you:

Recommended

Claude Fable 5.2·effort high

Agent
Claude Code
Subscription
Claude Max ($100) or Ultra ($200): Fable may use up to 50% of the weekly limit (on Pro, paid credits only)
Provider
Anthropic

≈ 2.9 agent-hours per 5-hour window · ≈ 14 per week on Claude Max 5x (central estimate). Simulate your case

Recommended

Claude Opus 5.5(default)·effort xhigh

Agent
Claude Code
Subscription
Claude Pro ($20), Max ($100) or Ultra ($200)
Provider
Anthropic

≈ 1.7 agent-hours per 5-hour window · ≈ 13 per week on Claude Pro (central estimate). Simulate your case

How long, then what

Stay on it
Only for the plan: one session, until the plan or SPEC.md is written, reviewed and approved.
Then
Stop there. Implement in a fresh session with Opus 5.5 at medium (default) — see Everyday coding — or hand the plan to a long run with /goal. Staying at xhigh to implement costs 2.5× more per task (AA index: $6.05 vs $2.45), and implementing with Fable 5.2 would spend its weekly cap on the part a cheaper model does well.
Alternative

GPT-6 Astra·effort xhigh

Codex · ChatGPT Plus ($20) or Pro ($100 / $200 / $500)

Codex: 55.3 on the AA index at $3.66; max adds only +1.5 points for +70% cost.

Alternative

Gemini 4 Carbon·effort high(default)

Antigravity · Google AI Pro ($19.99) or Ultra ($99.99 / $199.99)

Antigravity on Google AI Ultra: 60.0 on the AA index at high for $7.76 per task (at list price; half until 2026-11-30), above Opus 5.5 at xhigh (57.4), with few hallucinations. On Google AI Pro, Gemini 4 Argon at high: 50.4 on the AA index and first on LiveBench 2.0 reasoning (95.1, against 92–93 for Astra, GPT-6.1 Sol and Opus 5.5, one run); one planning session fits Pro's window.

How to work

  1. Type /plan <description> and set /effort xhigh (Opus 5.5), or /model fable then /effort high on Max or Ultra (Claude Code sets Fable 5.2 to medium by default): the agent reads and proposes, nothing is edited until you approve.
  2. End the spec with an end-to-end check that proves the feature works, then have a fresh session or subagent review the plan before implementing.
  3. For a large feature, ask Claude to interview you and write the result to SPEC.md; plans are also saved in ~/.claude/plans.
  4. Edit the plan with Ctrl+G, then implement in a fresh session at /effort medium.
  5. In Codex: /plan, with plan_mode_reasoning_effort set in config.toml, and keep a PLANS.md referenced from AGENTS.md.

Why

  • A plan is one long reasoning pass, so effort pays off: Opus 5.5 scores 57.4 on the AA Intelligence Index (v5) at xhigh, above Claude Sonnet 5.5 at max (54.5 for $10.02) and just below its own max (59.1), for 40% less than max ($6.05 vs $10.04 per task). It works on Claude Pro.
  • Planning is where Claude Fable 5.2 earns its quota on Max and Ultra: first on the AA index (61.8 at high for $9.08; 64.6 at max), first on FrontierCode 2.0 at xhigh (49.1%; 45.3% at high, level with Opus 5.5 at max, 45.5%, for less than its cost) and on LiveBench 2.0 overall (84.9). A plan is short in tokens but decides everything after it, so a few Fable turns pay off; then hand the work to Opus 5.5, which costs about 1.4 times less quota per hour.
  • Opus 5.5 is still first on LibraryDesignBench (48.9 ±0.9), which grades the API a model designs for a library, ahead of Claude Fable 5.1 (47.5) and GPT-6 Astra (45.1) — and of the real production library (46.6); Fable 5.2 is not ranked there yet.
  • Plan once at high (Fable 5.2) or xhigh (Opus 5.5), max only for the plan that really matters, then execute the steps at medium (Opus 5.5). Fable 5.2 at max is rarely worth it: +2.7 points on the AA index over high for nearly 3 times the cost ($26.54 vs $9.08).

Sources: Artificial Analysis model leaderboard · Snorkel LibraryDesignBench 1.0.0 · CWE-bench v2 · LiveBench · FrontierCode leaderboard data · Introducing Claude Fable 5.2 and Claude Mythos 5.2 · Claude Fable models on your plan · Gemini 4 Carbon is now available · Antigravity models · Claude Code best practices · Codex best practices

Habits that pay off in every agent

Whatever the task: what the Claude Code, Codex, Antigravity and Grok Build docs recommend, checked against their current versions (outdated tips left out).

  1. 01Verification first

    Give the agent a pass/fail check (tests, build, screenshot) and ask for evidence, not claims. Every vendor puts this tip first.

    Sources: Claude Code best practices · Codex best practices · Antigravity CLI best practices

  2. 02Plan only when it helps

    Plan ambiguous, multi-file or unfamiliar work; skip it for one-sentence changes (Claude Code and Grok Build docs). /plan is there when you need it.

    Sources: Claude Code best practices · Grok Build plan mode

  3. 03One task per session

    /clear between unrelated tasks, /rewind instead of piling up corrections, /btw for a side question, /compact <what to keep> mid-task.

    Sources: Claude Code best practices · Claude Code commands

  4. 04Short instruction files

    Keep AGENTS.md / CLAUDE.md short: add a rule when a mistake repeats, remove what the model already does, and re-test after each new model (/doctor).

    Sources: Claude Code best practices · Codex best practices

  5. 05Raise the effort, don't beg

    Start at the agent's default effort; raise it (or ultrathink for one turn) instead of writing “think harder”, and keep xhigh and max for tasks where they measurably help.

    Sources: Prompting Claude Opus 5.5 · Model configuration

  6. 06Fresh-context review

    Have code checked by an agent that did not write it (/code-review, /review), and ask it to report only correctness problems.

    Sources: Claude Code best practices · Claude Code commands

  7. 07Parallel work costs quota

    Worktrees, parallel sessions, subagents and workflows multiply the work done, but all draw on the same subscription limits: check the simulator before scaling out.

    Sources: Claude Code costs · Claude Code worktrees

  8. 08Auto mode, not skip-permissions

    Use auto mode (a classifier screens each action) or a sandbox; never --dangerously-skip-permissions in a normal session.

    Sources: Claude Code permission modes

guide/workflows

Practical coding-agent workflows

Choose a task for its method, commands for your agent and a request you can adapt. Each command names its surface; availability depends on your version and enabled skills.

What do you want to do?

Your first task

Use this for a small, visible change you can check yourself.

Before you start. A working starting point, one desired result, and a saved copy or Git commit.

  1. Open the relevant project and confirm the agent can see the right files.
  2. Describe one change and its acceptance criteria. Ask for a short plan if the scope is unclear.
  3. Inspect the diff, run the relevant checks, then try the result yourself.

Useful commands

/diffClaude Code CLI · Codex CLI
Inspect changes before accepting them.
/runClaude Code CLI
Launch and operate the app to observe the result.

Command documentation: Claude Code commands (2026-09-29) · Codex developer commands (2026-09-29) · Claude Code: run and verify your app (2026-09-29)

Request to adapt

In [project], implement [one change]. Preserve [constraints]. Success means [observable result]. First inspect the relevant files. Run [checks] and report the result, changed files and remaining limitations.

Done when. For a form: submit valid and invalid values, reload and verify persistence if required, then try a narrow screen.

If it fails. If the scope grows, save the current state and split the next change into a separate task.

Sources: Codex best practices (2026-10-10) · Claude Code best practices (2026-09-28)

guide/what-not-to-do

What not to do right now

The habits that waste quota or quality with today’s models and plans — each with what to do instead and the figures behind it. Re-checked at every data update.

Models

Reasoning effort

Habits

  • Don’t: Turn on /fast with Claude Fable 5.2 on a subscription

    Instead: Normal speed within your plan; fast mode only when you accept paying for it.

    Fable 5.2's fast mode (up to 2.5 times faster output) is billed at $16 / $80 per million tokens, twice the normal price, and only as paid usage credits on Claude plans: it never draws on the included quota.

    Sources: Claude Fable 5.2 model page · Claude Fable models on your plan

  • Don’t: Run ultracode, a workflow or an unbounded /swarm for ordinary tasks

    Instead: One session for normal work; keep workflows for a large migration or audit, tried on one directory first.

    A workflow orchestrates dozens of agents, and every agent draws on the same subscription limits: Claude Code warns beyond 25 agents or 1.5M tokens, and Google warns that Antigravity's multi-agent /swarm uses quota much faster. Combined with Fable or max, a single run can burn a week of quota.

    Sources: Claude Code workflows · Claude Code costs · Antigravity plans

  • Don’t: Run with --dangerously-skip-permissions to go faster

    Instead: Auto mode (the default): a classifier screens each action; or a sandbox or container.

    It removes every safeguard on file deletion, shell commands and network access, while auto mode already avoids most prompts.

    Sources: Claude Code permission modes · Claude Code best practices

  • Don’t: Force plan mode on every small change

    Instead: Plan only ambiguous or multi-file work with /plan.

    Claude Code's and Grok Build's docs both say to skip the plan when the change is obvious or fits in one sentence: for small changes it mostly costs a round trip.

    Sources: Claude Code best practices · Grok Build plan mode

quota/simulator

How long will your subscriptions last?

Enter the subscriptions you have and how you work: you get the agent hours per 5-hour window and per week, and when the quota runs out.

benchmarks/performance

Performance against quota burn

Each line is one model across its reasoning-effort levels. Further left = the task burns less of your quota; higher = better result. The dashed red step line is the efficient frontier: no other model × effort does better for less.

Benchmark

Ordered by confidence (⭐⭐⭐ = independent, controlled, coding-specific and current), then by the most cited.

Confidence 3 out of 3 stars of confidenceIndependent, same harness for every model at every effort level, more than 100 models; the widest controlled run of the reference agentic terminal benchmark.

IndependentAgentic terminal tasks run by Artificial Analysis with the same harness for every model, at every effort level.

Source updated continuously · Retrieved Nov 11, 2026

Score vs cost per task

API-equivalent cost at list prices, log scale — a secondary signal of relative consumption; subscription quota rules differ.

Scroll or pinch to zoom (Shift + scroll for the vertical axis), drag the slider, or use the zoom tool at the top right. Click a legend entry to hide or show it.

Loading the results…

Score by reasoning effort

Where the line flattens, the next effort level costs more quota for little or no gain.

Loading the results…

Show the data table

benchmarks/leaderboards

Coding leaderboards

Independent rankings of coding agents and web development. Switch each leaderboard on or off; scroll or zoom inside a chart, and click a provider in its legend to hide it.

Ordered by confidence (⭐⭐⭐ = independent, controlled, coding-focused and current; ⭐ = self-reported aggregates or stale sources), then by how widely the board is cited.

Arena Code: WebDev

Blind pairwise human votes on web-development tasks (995,980 votes, cutoff 2026-11-10).

Confidence 3 out of 3 stars of confidenceThe most cited coding arena: a very large number of blind human votes on real web apps, updated continuously.

Useful for: Front-end and web UI · 3D, games and creative coding as code

Source updated Nov 10, 2026 · Retrieved Nov 11, 2026 · 157 models

Anomaly detected: Claude Fable 5.2 (xHigh) enters first on only 994 votes (±22 points), and Gemini 4 Carbon (High) is fourth as a pre-release model. On 2026-11-10 Arena replaced the proprietary Mistral Large 4 entry with the open-weight model under the same name, so its two ratings are not comparable.

Loading the rows…

Show the data table

Artificial Analysis Coding Agent Index

Each coding agent with its vendor's model (Claude Code, Codex, Grok Build, Kimi Code, Antigravity, OpenCode, Muse Code, Devin Fusion, GitHub Copilot, Mistral Vibe, Mirror, DeepSeek Harness) on FrontierCode v1.1, Terminal-Bench 4.0 and SWE-Atlas-QnA — the closest proxy for what a subscription gives you.

Confidence 3 out of 3 stars of confidenceIndependent, measures the agents themselves (Claude Code, Codex, Antigravity…) with their models, on current versions.

Useful for: Everyday coding · Long autonomous agent runs · High volume, subagents, quick edits

Source updated Nov 10, 2026 · Retrieved Nov 11, 2026 · 35 models

Incomplete coverage — this data does not include Claude Fable 5.2, Kimi K3.1 Pro, MiMo-V2.6-Pro. Key current models are missing, so the ranking here is not fully relevant.

Anomaly detected: On 2026-10-27 Artificial Analysis replaced DeepSWE v1.1 with FrontierCode v1.1 in the index and re-scored SWE-Atlas-QnA: scores are not comparable with earlier ones. GPT-6.1 Terra scores the same at xhigh and max (48.2 and 48.1) for 56% of the cost at xhigh. Gemini 4 Argon's cost doubled on 2026-11-10 with an unchanged score, after AA doubled its API price. Muse Wonder Full and Muse Ember have no published cost: their cost per task is missing, not zero. Claude Fable 5.2 has no row yet.

Loading the rows…

Show the data table

Scale SWE-Bench Pro V2 (hard)

Real issues of commercial and open-source repositories, re-graded on a clean image, each model in its vendor's agent (Claude Code, Codex) or mini-swe-agent. The hard subset: the full set sits near 99% for frontier models.

Confidence 3 out of 3 stars of confidenceIndependent, real issues of commercial and open-source repositories re-graded on a clean image, current models.

Useful for: Debugging · Everyday coding

Source updated Nov 6, 2026 · Retrieved Nov 11, 2026 · 18 models

Incomplete coverage — this data does not include Claude Opus 5.5, Claude Fable 5.2, GPT-6.1 Sol, GPT-6.1 Terra, GPT-6 Luna, Gemini 4 Carbon, Grok 4.7, Kimi K3.1 Pro, Qwen3.8 Max, MiMo-V2.6-Pro. Key current models are missing, so the ranking here is not fully relevant.

Anomaly detected: Claude Sonnet 5.5 (mini-swe-agent, xhigh) scores 70.6% on the hard subset against 95.8% on the full set, a far larger drop than any other frontier model (Gemini 4 Argon 92.2 against 99.2). The hard subset publishes no confidence interval (0) except for Claude Haiku 5.5 (±13.0), and its 51 tasks move scores in steps of about 2 points: compare ranks, not small gaps. Inkling 2 Small (added 2026-11-06) is listed without a harness or an effort.

Loading the rows…

Show the data table

LiveBench 2.0: long-horizon

LiveBench 2.0 (release of 2026-11-06, updated 2026-11-09): 18 long-running agentic tasks in six groups (apps extended over several turns of changing requirements, delegated and long-run operations on a codebase over simulated weeks, systems and performance work), deterministic grading, each model at its highest supported effort unless its name says otherwise, with the cost per task.

Confidence 3 out of 3 stars of confidenceIndependent, contamination-limited long tasks, regularly refreshed, current models.

Useful for: Long autonomous agent runs · Everyday coding

Source updated Nov 9, 2026 · Retrieved Nov 11, 2026 · 33 models

Incomplete coverage — this data does not include Gemini 4 Carbon, GLM-5.3. Key current models are missing, so the ranking here is not fully relevant.

Anomaly detected: Each suite is a single run: between the 2026-10-25 and 2026-11-06 releases the same tasks moved some models by up to 22 points (Muse Spark 1.3 25.5 → 47.9), so read small gaps as noise. Claude Fable 5.2 is shown at xhigh because its max run scored lower overall; its cost per task ($160.93) is about twice Claude Opus 5.5's. LiveBench rewrote its earlier releases after publication (costs re-measured, models added). Mistral Large 4 appears twice: as "Mistral Large 4 1029" (24.7) and under its bare name (9.2).

Loading the rows…

Show the data table

Arena Agent Arena: code

Real coding sessions of Arena users with each model in an agent, judged on signals such as confirmed success, praise and complaints, steering effort and tool errors (909,397 code sessions). Live user sessions, not a fixed task set.

Confidence 3 out of 3 stars of confidenceReal coding-agent sessions rated by their users, close to a million sessions, current models.

Useful for: Everyday coding · Long autonomous agent runs

Source updated Nov 8, 2026 · Retrieved Nov 11, 2026 · 57 models

Incomplete coverage — this data does not include Claude Fable 5.2. Key current models are missing, so the ranking here is not fully relevant.

Anomaly detected: Arena changed two of the five signals behind this score: Steerability on 2026-10-01 (every judged turn instead of the turn after a correction) and Bash Recovery, replaced by Bash Reliability on 2026-10-12 (bash failures per call, detected by an LLM judge). Snapshots before 2026-10-16 are not comparable with this one. On 2026-11-08 four rows (DeepSeek-V4.1-Flash-20260910, DeepSeek-V4-Pro High, MiMo-V2.5-Pro and the proprietary Mistral Large 4) were dropped without explanation, and Mistral Large 4 now refers to the open-weight model.

Loading the rows…

Show the data table

FrontierSWE V2

34 very hard engineering tasks with a 20-hour budget each — the long-horizon extreme.

Confidence 3 out of 3 stars of confidenceIndependent, hard software-engineering tasks scored as the mean of five trials, current models.

Useful for: Long autonomous agent runs

Source updated Nov 9, 2026 · Retrieved Nov 11, 2026 · 24 models

Incomplete coverage — this data does not include GPT-6.1 Sol, GPT-6.1 Terra, GPT-6 Luna, Kimi K3.1 Pro, MiMo-V2.6-Pro. Key current models are missing, so the ranking here is not fully relevant.

Anomaly detected: The source warns that Gemini 4 Argon's cost ($129.36 per trial) was measured with a much lower cache hit rate than in production (a non-production setting): its cost is overstated compared with real use.

Loading the rows…

Show the data table

SWE-Marathon v1.1

20 multi-hour engineering tasks, 8 trials each, every model in an agent (mostly Claude Code) at the published effort, with the cost of each trial.

Confidence 3 out of 3 stars of confidenceIndependent long tasks with many trials per model, recent and covering current models.

Useful for: Long autonomous agent runs

Source updated Nov 7, 2026 · Retrieved Nov 11, 2026 · 26 models

Incomplete coverage — this data does not include Claude Fable 5.2, GPT-6.1 Terra, Gemini 3.8 Flash, Kimi K3.1 Pro, Qwen3.8 Max, MiMo-V2.6-Pro. Key current models are missing, so the ranking here is not fully relevant.

Anomaly detected: On 2026-10-15 the site removed 11 older models from its board (Claude Opus 5, Opus 4.8, Sonnet 5, Fable 5, GPT-6 Sol, GPT-5.6 Sol/Terra/Luna, Grok 4.5, Grok 4.6, GLM 5.2); their results are unchanged at the source but, as on the site, they are no longer shown here. On 2026-10-20 it replaced Muse Spark 1.3's September runs with a new max run (46.25%, $14.29 per trial). Gemini 4 Carbon (Antigravity CLI, high, 53.75%), Muse Wonder Full and Muse Wonder lead with no published cost. Inkling 2 Small costs $424 per trial. Kimi K3 costs $661 per trial (133 M tokens), 4 to 10 times any other run: likely a pricing or token-counting problem at the source. Claude Haiku 5.5 at max (50.0%) is above Claude Opus 5.5 at max (46.25%), a gap of 6 trials out of 160 that is within noise.

Loading the rows…

Show the data table

LiveBench v1: agentic coding

LiveBench v1 (release 2026-06-25, now archived by LiveBench 2.0): contamination-limited questions; agentic coding in JavaScript, TypeScript and Python.

Confidence 2 out of 3 stars of confidenceIndependent and respected, but the v1 suite is now superseded by LiveBench 2.0.

Useful for: Everyday coding · High volume, subagents, quick edits

Source updated Oct 16, 2026 · Retrieved Nov 11, 2026 · 71 models

Incomplete coverage — this data does not include Claude Fable 5.2, GPT-6.1 Terra, Gemini 4 Carbon, Gemini 4 Argon, Kimi K3.1 Pro, MiMo-V2.6-Pro. Key current models are missing, so the ranking here is not fully relevant.

Loading the rows…

Show the data table

Design Arena: code overall

All of Design Arena's code categories in one ranking (1,950,799 votes): websites, components, games, SVG, charts and more. No standard error is published, so close Elo scores cannot be told apart.

Confidence 2 out of 3 stars of confidenceBlind human votes on generated code, independent, but fewer votes than Arena and judged on looks first.

Useful for: Front-end and web UI · SVG illustrations and icons as code · 3D, games and creative coding as code

Source updated Nov 11, 2026 · Retrieved Nov 11, 2026 · 195 models

Incomplete coverage — this data does not include Gemini 4 Carbon. Key current models are missing, so the ranking here is not fully relevant.

Loading the rows…

Show the data table

Design Arena: agentic web app, full-stack

Two agents build the same full-stack app (React, Supabase, deployed to Vercel) in a sandbox with 23 tools; people vote for the better one.

Confidence 2 out of 3 stars of confidenceBlind human votes on agent-built full-stack apps; independent, but a smaller and design-led vote.

Useful for: Front-end and web UI · Long autonomous agent runs

Source updated Nov 11, 2026 · Retrieved Nov 11, 2026 · 51 models

Incomplete coverage — this data does not include Claude Fable 5.2, Claude Sonnet 5.5, Claude Haiku 5.5, GPT-6.1 Sol, GPT-6.1 Terra, Gemini 4 Carbon, Gemini 4 Argon, Grok 4.7, GLM-5.3, MiMo-V2.6-Pro. Key current models are missing, so the ranking here is not fully relevant.

Anomaly detected: Design Arena has published no standard error since 2026-10-10 (empty for every row of all its boards), so no confidence interval is shown: compare ranks, not small Elo gaps. On 2026-11-11 Claude Opus 5.5 entered first on only 318 battles, and every other model lost about 19 Elo at once.

Loading the rows…

Show the data table

Design Arena: agentic web app, frontend

Same arena for multi-file React front-ends (13 tools).

Confidence 2 out of 3 stars of confidenceBlind human votes on agent-built front ends; independent, but a smaller and design-led vote.

Useful for: Front-end and web UI

Source updated Nov 11, 2026 · Retrieved Nov 11, 2026 · 56 models

Incomplete coverage — this data does not include Claude Fable 5.2, Claude Haiku 5.5, GPT-6.1 Terra, Gemini 4 Carbon, Gemini 4 Argon, Grok 4.7. Key current models are missing, so the ranking here is not fully relevant.

Anomaly detected: Design Arena has published no standard error since 2026-10-10, so no confidence interval is shown. GPT-6.1 Sol, fifth on 523 battles on 2026-10-27, fell to 14th (−39 Elo) on 871: early ranks on this board move a lot.

Loading the rows…

Show the data table

Design Arena: website

Single-shot websites generated by each model, judged by blind votes (1,190,870 votes, Design Arena's largest web board). No standard error is published, so close Elo scores cannot be told apart.

Confidence 2 out of 3 stars of confidenceBlind human votes on websites; independent, but narrow and judged on looks.

Useful for: Front-end and web UI

Source updated Nov 11, 2026 · Retrieved Nov 11, 2026 · 206 models

Incomplete coverage — this data does not include Gemini 4 Carbon. Key current models are missing, so the ranking here is not fully relevant.

Loading the rows…

Show the data table

Arena Code: Image-to-WebDev

Blind pairwise votes on web pages built from an image, such as a mockup or a screenshot (210,832 votes, cutoff 2026-11-10).

Confidence 2 out of 3 stars of confidenceArena's blind votes, but narrow: turning a screenshot into a web page.

Useful for: Front-end and web UI

Source updated Nov 10, 2026 · Retrieved Nov 11, 2026 · 66 models

Incomplete coverage — this data does not include Claude Fable 5.2, Claude Haiku 5.5, GPT-6.1 Terra, Gemini 4 Carbon, Gemini 4 Argon, Kimi K3.1 Pro, GLM-5.3. Key current models are missing, so the ranking here is not fully relevant.

Loading the rows…

Show the data table

WeirdML v3 (agentic)

Unusual machine-learning problems solved end to end by an agent (Claude Code, Codex CLI, Gemini CLI, Kimi Code, OpenCode), with the cost of each run. Data from Epoch AI.

Confidence 2 out of 3 stars of confidenceIndependent and repeated runs, but machine-learning tasks rather than everyday software work.

Useful for: Algorithms, proofs and competitive programming · Scientific and numerical code

Source updated Nov 11, 2026 · Retrieved Nov 11, 2026 · 26 models

Incomplete coverage — this data does not include Claude Fable 5.2, Gemini 4 Carbon, GLM-5.3, Qwen3.8 Max, MiMo-V2.6-Pro. Key current models are missing, so the ranking here is not fully relevant.

Loading the rows…

Show the data table

GSO (software optimization)

Making real repositories faster without breaking them, in OpenHands, scored after removing reward hacks. Only GPT-6 Astra among current models.

Confidence 2 out of 3 stars of confidenceIndependent, hack-adjusted, but narrow (performance optimization) and updated less often.

Useful for: Algorithms, proofs and competitive programming

Source updated Sep 27, 2026 · Retrieved Oct 11, 2026 · 32 models

Incomplete coverage — this data does not include Claude Opus 5.5, Claude Fable 5.2, Claude Sonnet 5.5, Claude Haiku 5.5, GPT-6.1 Sol, GPT-6.1 Terra, GPT-6 Luna, Gemini 4 Carbon, Gemini 4 Argon, Gemini 3.8 Flash, Grok 4.7, Kimi K3.1 Pro, Kimi K3, GLM-5.3, Qwen3.8 Max, MiMo-V2.6-Pro. Key current models are missing, so the ranking here is not fully relevant.

Anomaly detected: Rows dated 2026-09-27 use a new elicitation and jump to 70–88%: they are not comparable with older rows.

Loading the rows…

Show the data table

Scale SWE Atlas: codebase Q&A

Answering deep questions about a real codebase by exploring and running it, in the vendor's agent — one third of the Artificial Analysis Coding Agent Index.

Confidence 2 out of 3 stars of confidenceIndependent (Scale), but a narrow skill (answering questions on a codebase) and not refreshed recently.

Useful for: Advanced planning and architecture · Everyday coding

Source updated Sep 9, 2026 · Retrieved Sep 29, 2026 · 24 models

Incomplete coverage — this data does not include Claude Opus 5.5, Claude Fable 5.2, Claude Sonnet 5.5, Claude Haiku 5.5, GPT-6.1 Sol, GPT-6.1 Terra, GPT-6 Luna, Gemini 4 Carbon, Gemini 4 Argon, Grok 4.7, Kimi K3.1 Pro, Kimi K3, GLM-5.3, Qwen3.8 Max, MiMo-V2.6-Pro. Key current models are missing, so the ranking here is not fully relevant.

Loading the rows…

Show the data table

Scale SWE Atlas: test writing

Writing tests that catch real bugs in an existing codebase, in the vendor's agent.

Confidence 2 out of 3 stars of confidenceIndependent (Scale), but a narrow skill (writing tests) and not refreshed recently.

Useful for: Everyday coding · High volume, subagents, quick edits

Source updated Sep 9, 2026 · Retrieved Sep 29, 2026 · 24 models

Incomplete coverage — this data does not include Claude Opus 5.5, Claude Fable 5.2, Claude Sonnet 5.5, Claude Haiku 5.5, GPT-6.1 Sol, GPT-6.1 Terra, GPT-6 Luna, Gemini 4 Carbon, Gemini 4 Argon, Grok 4.7, Kimi K3.1 Pro, Kimi K3, GLM-5.3, Qwen3.8 Max, MiMo-V2.6-Pro. Key current models are missing, so the ranking here is not fully relevant.

Loading the rows…

Show the data table

Scale SWE Atlas: refactoring

Refactoring real code without breaking its behavior, in the vendor's agent.

Confidence 2 out of 3 stars of confidenceIndependent (Scale), but a narrow skill (refactoring) and not refreshed recently.

Useful for: High volume, subagents, quick edits · Everyday coding

Source updated Sep 9, 2026 · Retrieved Sep 29, 2026 · 17 models

Incomplete coverage — this data does not include Claude Opus 5.5, Claude Fable 5.2, Claude Sonnet 5.5, Claude Haiku 5.5, GPT-6.1 Sol, GPT-6.1 Terra, GPT-6 Luna, Gemini 4 Carbon, Gemini 4 Argon, Grok 4.7, Kimi K3.1 Pro, Kimi K3, GLM-5.3, Qwen3.8 Max, MiMo-V2.6-Pro. Key current models are missing, so the ranking here is not fully relevant.

Loading the rows…

Show the data table

Snorkel Agentic Coding 2.0

Terminal-style agentic tasks without network access, 5 attempts each, with the cost per trial. Run by Snorkel AI (a data vendor); no effort stated.

Confidence 2 out of 3 stars of confidenceIndependent agentic coding tasks with costs, but no effort stated and fewer models.

Useful for: Long autonomous agent runs · Everyday coding

Source updated Oct 23, 2026 · Retrieved Nov 11, 2026 · 16 models

Incomplete coverage — this data does not include Claude Fable 5.2, Claude Sonnet 5.5, Claude Haiku 5.5, GPT-6.1 Terra, GPT-6 Luna, Gemini 4 Carbon, Kimi K3.1 Pro, MiMo-V2.6-Pro. Key current models are missing, so the ranking here is not fully relevant.

Loading the rows…

Show the data table

Snorkel SWE-bench CLI

Repository changes that span at least two files, solved from the command line, with the cost per trial. Run by Snorkel AI; no effort stated.

Confidence 2 out of 3 stars of confidenceIndependent multi-file repository changes, but no effort stated and fewer models.

Useful for: Debugging · Everyday coding

Source updated Oct 23, 2026 · Retrieved Nov 11, 2026 · 16 models

Incomplete coverage — this data does not include Claude Fable 5.2, Claude Sonnet 5.5, Claude Haiku 5.5, GPT-6.1 Terra, GPT-6 Luna, Gemini 4 Carbon, Kimi K3.1 Pro, MiMo-V2.6-Pro. Key current models are missing, so the ranking here is not fully relevant.

Loading the rows…

Show the data table

Vals AI: ProgramBench

Rebuilding real programs from scratch, run by Vals AI with one effort per model, with the cost per test.

Confidence 2 out of 3 stars of confidenceIndependent, building a program from scratch, but one effort per model and a narrow task.

Useful for: Long autonomous agent runs · Advanced planning and architecture

Source updated Nov 9, 2026 · Retrieved Nov 11, 2026 · 66 models

Incomplete coverage — this data does not include GPT-6.1 Terra, Gemini 4 Carbon, Kimi K3.1 Pro. Key current models are missing, so the ranking here is not fully relevant.

Loading the rows…

Show the data table

Vals AI: SRE-bench

Reverse-engineering real, contamination-free binaries, run by Vals AI with one effort per model, with the cost per test.

Confidence 2 out of 3 stars of confidenceIndependent and contamination-free, but a narrow skill (reverse engineering) and one effort per model.

Useful for: Long autonomous agent runs

Source updated Nov 5, 2026 · Retrieved Nov 11, 2026 · 37 models

Incomplete coverage — this data does not include Claude Fable 5.2, GPT-6.1 Terra, Gemini 4 Carbon, Kimi K3, Qwen3.8 Max. Key current models are missing, so the ranking here is not fully relevant.

Loading the rows…

Show the data table

CWE-bench v2

Finding and fixing real vulnerabilities in real codebases (250 audit-and-patch tasks across 185 CWEs: 100 public, 150 held out), each model in its own agent at high reasoning, graded by deterministic verifiers (pass@1 over 4 rollouts), with the mean cost per rollout. Run by Collinear AI.

Confidence 2 out of 3 stars of confidenceIndependent security-fix tasks, recent, but narrow (vulnerabilities).

Useful for: Everyday coding

Source updated Nov 2, 2026 · Retrieved Nov 11, 2026 · 11 models

Incomplete coverage — this data does not include Claude Fable 5.2, Claude Sonnet 5.5, Claude Haiku 5.5, GPT-6.1 Sol, GPT-6.1 Terra, GPT-6 Luna, Gemini 4 Carbon, Gemini 3.8 Flash, Kimi K3, MiMo-V2.6-Pro. Key current models are missing, so the ranking here is not fully relevant.

Anomaly detected: CWE-bench v2 (2026-11-02) replaced v1 with more tasks, more weakness types and stricter verifiers, so v1 scores do not compare. The top four (Claude Opus 5.5 67.1%, Muse Wonder 66.1%, Gemini 4 Argon 65.0%, GPT-6 Astra 64.2%) are within the ±5-point margin. Claude Fable 5.2 and Gemini 4 Carbon are not measured on v2.

Loading the rows…

Show the data table

Endor Labs Agent Security League

Does the code an agent writes pass both functional and security tests? Share of tasks that are correct and secure, per agent × model (Claude Code, Codex, Cursor, Muse Code). Run by Endor Labs (a security vendor).

Confidence 2 out of 3 stars of confidenceIndependent, secure and correct code, but narrow (security) and fewer models.

Useful for: Everyday coding

Source updated Oct 24, 2026 · Retrieved Nov 11, 2026 · 34 models

Incomplete coverage — this data does not include Claude Fable 5.2, Claude Sonnet 5.5, Claude Haiku 5.5, GPT-6.1 Terra, GPT-6 Luna, Gemini 4 Carbon, Gemini 3.8 Flash, Grok 4.7, Kimi K3.1 Pro, Kimi K3, GLM-5.3, Qwen3.8 Max, MiMo-V2.6-Pro. Key current models are missing, so the ranking here is not fully relevant.

Anomaly detected: Between 2026-10-16 and 2026-10-24 Endor Labs re-scored almost every row without a note, mostly downward (Claude Code + Claude Opus 5.5 33.5% → 30.7% secure), after adding anti-cheating safeguards to its method. Claude Opus 5.5 remains less often functionally correct than Claude Opus 5 (64.8% against 73.7%) and is now also less often secure (30.7% against 32.4%).

Loading the rows…

Show the data table

Terminal-Bench-Science (Artificial Analysis run)

Artificial Analysis's own run of Terminal-Bench-Science with the same harness for every model, at several efforts (mean of 3 repeats).

Confidence 2 out of 3 stars of confidenceIndependent run with the same harness for every model, but scientific terminal tasks rather than software work.

Useful for: Scientific and numerical code

Source updated Nov 9, 2026 · Retrieved Nov 11, 2026 · 135 models

Anomaly detected: Claude rows were run with a default fallback model, and AA ranks Claude Opus 5.5 at xhigh (61.9%) above max (59.05%), while the Snorkel mirror gives Opus 5.5 max 63.3% in Claude Code. Claude Fable 5.2 scores higher at high (70.0%) than at xhigh (67.6%). GPT-6.1 Terra scores higher without reasoning (24.3%) than at low (7.6%).

Loading the rows…

Show the data table

Design Arena: UI components

Blind votes on interface components written as code (114,714 votes). No standard error is published, so close Elo scores cannot be told apart.

Confidence 2 out of 3 stars of confidenceBlind human votes on UI components; independent, but narrow and judged on looks.

Useful for: Front-end and web UI

Source updated Nov 11, 2026 · Retrieved Nov 11, 2026 · 186 models

Incomplete coverage — this data does not include Gemini 4 Carbon. Key current models are missing, so the ranking here is not fully relevant.

Anomaly detected: The top two are 6 Elo apart: GPT-6 Astra (xhigh) on 669 battles and Kimi K3.1 Pro on 324. Claude Fable 5.2 entered 16th on only 125: treat these ranks as provisional.

Loading the rows…

Show the data table

Design Arena: game dev

Blind votes on browser games written as code (318,370 votes) — the closest board to 3D and creative coding. No standard error is published, so close Elo scores cannot be told apart.

Confidence 2 out of 3 stars of confidenceBlind human votes on small games; independent, but narrow.

Useful for: 3D, games and creative coding as code

Source updated Nov 11, 2026 · Retrieved Nov 11, 2026 · 193 models

Incomplete coverage — this data does not include Gemini 4 Carbon. Key current models are missing, so the ranking here is not fully relevant.

Anomaly detected: Claude Sonnet 5.5, first on 435 battles on 2026-10-16, is now ninth on 2,008: early ranks on this board move a lot. Claude Fable 5.2 entered eighth on 396 battles.

Loading the rows…

Show the data table

Design Arena: mobile apps (agents)

Two agents build the same mobile app and people vote for the better one (45,185 votes): it compares agents, not bare models. No standard error is published, so close Elo scores cannot be told apart.

Confidence 2 out of 3 stars of confidenceBlind human votes on agent-built mobile apps; independent, but narrow and a smaller vote.

Useful for: Front-end and web UI

Source updated Nov 11, 2026 · Retrieved Nov 11, 2026 · 58 models

Incomplete coverage — this data does not include Claude Haiku 5.5, Gemini 4 Carbon, Gemini 4 Argon, MiMo-V2.6-Pro. Key current models are missing, so the ranking here is not fully relevant.

Anomaly detected: Claude Fable 5.2 enters first on only 179 battles, 40 Elo ahead of Claude Opus 5.5, and no standard error is published: treat its rank as provisional.

Loading the rows…

Show the data table

Design Arena: SVG

Blind votes on SVG icons, illustrations and logos written as code (211,612 votes). Includes SVG-specialized models that no subscription offers.

Confidence 2 out of 3 stars of confidenceBlind human votes on SVG drawings; independent, but a very narrow skill.

Useful for: SVG illustrations and icons as code

Source updated Nov 11, 2026 · Retrieved Nov 11, 2026 · 139 models

Incomplete coverage — this data does not include Claude Fable 5.2, GPT-6.1 Sol, GPT-6.1 Terra, Gemini 4 Carbon, Qwen3.8 Max. Key current models are missing, so the ranking here is not fully relevant.

Anomaly detected: Design Arena stopped publishing standard errors on 2026-10-10 (they were published for this board on 2026-09-29), so no confidence interval is shown: compare ranks, not small Elo gaps.

Loading the rows…

Show the data table

Design Arena: data visualization

Blind votes on charts and dashboards written as code (90,018 votes; no standard error published).

Confidence 2 out of 3 stars of confidenceBlind human votes on charts; independent, but a very narrow skill.

Useful for: Charts, dashboards and diagrams as code · Front-end and web UI

Source updated Nov 11, 2026 · Retrieved Nov 11, 2026 · 156 models

Incomplete coverage — this data does not include Claude Fable 5.2, Gemini 4 Carbon. Key current models are missing, so the ranking here is not fully relevant.

Anomaly detected: Since 2026-10-16 the board lists only models with at least 100 battles, a rule the site does not state. Claude Sonnet 5.5 is back second on 160 battles, and no standard error is published: treat its rank as provisional.

Loading the rows…

Show the data table

llm-stats: Best AI for coding

Aggregate rating across 68 coding benchmarks, mostly self-reported by vendors.

Confidence 1 out of 3 stars of confidenceAn aggregate that mixes many scores, most of them self-reported by the vendors: useful context, not a measurement.

Useful for: Everyday coding

Source updated Nov 11, 2026 · Retrieved Nov 11, 2026 · 207 models

Incomplete coverage — this data does not include Gemini 4 Carbon. Key current models are missing, so the ranking here is not fully relevant.

Anomaly detected: Muse Wonder Full and Muse Wonder are each listed twice on 2026-11-11, under two model ids with different ratings; the site gives no explanation. Claude Mythos 5.2 enters fourth on only two self-reported results, and Gemini 4 Carbon is in the index but hidden from the table.

Loading the rows…

Show the data table

SWE-rebenchstale

Fresh GitHub issues, decontaminated — but the last window ends on 2026-07-01, before most current models.

Confidence 1 out of 3 stars of confidenceRespected method, but its last window predates most current models.

Useful for: Debugging · Everyday coding

Source updated Jul 24, 2026 · Retrieved Sep 28, 2026 · 17 models

Incomplete coverage — this data does not include Claude Opus 5.5, Claude Fable 5.2, Claude Sonnet 5.5, Claude Haiku 5.5, GPT-6 Astra, GPT-6.1 Sol, GPT-6.1 Terra, GPT-6 Luna, Gemini 4 Carbon, Gemini 4 Argon, Gemini 3.8 Flash, Grok 4.7, Kimi K3.1 Pro, Kimi K3, GLM-5.3, Qwen3.8 Max, MiMo-V2.6-Pro. Key current models are missing, so the ranking here is not fully relevant.

Loading the rows…

Show the data table

KiloBench (Terminal-Bench 2.0 in Kilo)

Terminal-Bench 2.0 run in the Kilo agent; cost is per full run of the suite, not per task.

Confidence 1 out of 3 stars of confidenceRuns the older Terminal-Bench 2.0 in a single agent, and is not refreshed often.

Useful for: Everyday coding

Source updated Sep 28, 2026 · Retrieved Sep 28, 2026 · 38 models

Incomplete coverage — this data does not include Claude Opus 5.5, Claude Fable 5.2, Claude Sonnet 5.5, Claude Haiku 5.5, GPT-6.1 Sol, GPT-6.1 Terra, GPT-6 Luna, Gemini 4 Carbon, Gemini 4 Argon, Grok 4.7, Kimi K3.1 Pro, GLM-5.3, MiMo-V2.6-Pro. Key current models are missing, so the ranking here is not fully relevant.

Loading the rows…

Show the data table

Terminal-Bench-Science 0.1

Scientific work in a terminal (life, physical, earth, engineering and mathematical sciences) done by coding agents, each model in its vendor's agent at a stated effort. Official board, rows from Snorkel's mirror.

Confidence 1 out of 3 stars of confidenceVersion 0.1 mirror, superseded by the Artificial Analysis run of the same benchmark.

Useful for: Scientific and numerical code · Long autonomous agent runs

Source updated Sep 24, 2026 · Retrieved Nov 11, 2026 · 17 models

Incomplete coverage — this data does not include Claude Fable 5.2, Claude Sonnet 5.5, Claude Haiku 5.5, GPT-6.1 Sol, GPT-6.1 Terra, GPT-6 Luna, Gemini 4 Carbon, Gemini 4 Argon, Kimi K3.1 Pro, Qwen3.8 Max, MiMo-V2.6-Pro. Key current models are missing, so the ranking here is not fully relevant.

Loading the rows…

Show the data table

Snorkel LibraryDesignBench

Can the model design a library that other code then uses? Pass rate on downstream problems times the simplicity of the API, per model and harness. Run by Snorkel AI.

Confidence 1 out of 3 stars of confidenceNarrow (library API design), few models, not refreshed recently.

Useful for: Advanced planning and architecture

Source updated Oct 1, 2026 · Retrieved Nov 11, 2026 · 11 models

Incomplete coverage — this data does not include Claude Fable 5.2, Claude Sonnet 5.5, Claude Haiku 5.5, GPT-6.1 Sol, GPT-6.1 Terra, GPT-6 Luna, Gemini 4 Carbon, Gemini 4 Argon, Gemini 3.8 Flash, Grok 4.7, Kimi K3.1 Pro, Qwen3.8 Max, MiMo-V2.6-Pro. Key current models are missing, so the ranking here is not fully relevant.

Loading the rows…

Show the data table

benchmarks/vendor-figures

Official launch figures

What each vendor published for its own model at launch, with the effort it used. Vendors pick their own agent, harness and settings, so these numbers usually run higher than independent measurements — use them to see each vendor's claims, and the charts above to compare.

Incomplete coverage — this data does not include Qwen3.8 Max, MiMo-V2.6-Pro. Key current models are missing, so the ranking here is not fully relevant.

ModelSWE-bench VerifiedSWE-bench ProTerminal-Bench 2.xTerminal-Bench 4.0DeepSWE v1.1FrontierCode 2.0FrontierCode 1.1
Claude Fable 5.2Anthropic—93%max—67.6%max74.3%max49.1%xhigh—
Claude Fable 5.1Anthropic—81.2%max—55.8%max——50.3%max
Claude Opus 5.5Anthropic—89.9%max—66.4%xhigh74.2%max—54.4%max
Claude Sonnet 5Anthropic85.2%max63.2%max80.4%xhigh · 2.1————
Claude Sonnet 5.5Anthropic—81.3%max—70.6%max71%max—52.1%xhigh
Claude Haiku 5.5Anthropic—64.8%max—39.2%max——46.4%max
Claude Haiku 4.5Anthropic73.3%——————
GPT-6 AstraOpenAI———57.88%high74.12%xhigh—53.3%max
GPT-6.1 SolOpenAI————75.22%high——
GPT-6.1 TerraOpenAI———50.3%max———
GPT-6 SolOpenAI————68.81%max—49.27%max
GPT-6 LunaOpenAI————66.59%max—42.42%max
Gemini 4 CarbonGoogle———65.2%high (default)———
Gemini 4 ArgonGoogle———57.4%high (default)77.9%high (default)——
Gemini 3.8 FlashGoogle——89.4%medium · 2.119.1%73.7%high (default)——
Gemini 3.7 Flash (deprecated)Google——85.8%medium (default) · 2.111.2%65.3%——
Gemini 3.1 Pro (preview)Google80.6%high (default)54.2%high (default)68.5%high (default) · 2.0————
Grok 4.7xAI———38%xhigh71%high (default)——
Grok 4.6xAI———20.3%high (default)65.9%high (default)——
Kimi K3.1 ProMoonshot AI———65.7%max75.5%max——
Kimi K3Moonshot AI——88.3%max · 2.1—67.5%max——
GLM-5.3Z.ai——88.2%max (default) · 2.1—66.9%max (default)——
GLM-5.3-FlashZ.ai——84.3%2.1—63.4%——
Mistral Large 4mistral———41.4%61.7%preview19.7%high (default)—

landscape/remote-or-local

Remote or local?

When running a model on your own machine makes sense: hardware, performance and limits compared with remote agents.

Running models locally

Local models are a separate tier: they cost no quota, but the best of them reach 5–33% on Terminal-Bench 4.0 versus about 60% for frontier models. Pair a local model for routine work with a remote frontier model for hard tasks.

HardwareRecommended modelMemoryDecode speedQuality (AA index · Terminal-Bench 4.0)
24 GB GPU (RTX 4090 / 3090)Qwen3.8-27B Q4_K_M16.5 GB38–46 tok/sAA 33.7 · TB4 5.6%
32 GB GPU (RTX 5090)Qwen3.8-27B Q6 / NVFP422–25 GB23–75 tok/s · MTP 98–134AA 33.7 · TB4 5.6%
128 GB unified memory (Strix Halo, DGX Spark, M5 Max)Qwen3.8-Flash-Next IQ4_XS93.7 GB11–37 tok/s · MTP 47–84AA 39.8 · TB4 25.3%
256 GB unified memory (M5 Ultra, M3 Ultra)GLM-5.3-Flash MLX 4-bit204 GB26–41 tok/sAA 41.8 · TB4 32.8%

Worth it for

  • Private or regulated code that must not leave the machine
  • Working offline
  • Hardware you already own
  • High-volume routine work: boilerplate, tests, subagents
  • A 256 GB machine shared by a team

Not worth it for

  • Saving money versus a $20 subscription (hardware and power cost $10–270 a month)
  • Hard agentic tasks: the quality gap to frontier models is large

docs/faq

FAQ

Focused answers to choose a model, balance reasoning effort and manage subscription quotas.

What is an AI coding agent?

An AI that works inside a software project: it reads the files, edits the code, runs commands and tests, and iterates until the task is done, under the developer's control. The most used in 2026 are Claude Code, GitHub Copilot, Codex and Cursor.

Where and how an agent is used: a terminal (CLI), a desktop app, an IDE extension, a hosted web service, remotely from a phone or another browser, inside GitHub and CI/CD, from team tools (Slack, Teams, Linear, Jira) or through an SDK in your own scripts. The same agent often has several, each with its own name: Codex CLI, the ChatGPT desktop app, the Codex IDE extension, Codex cloud, Codex Remote, Codex code review and the Codex SDK.

Sources: AI coding agent adoption 2026 · Codex glossary · Details on the page

Which agent, model and effort for “Advanced planning and architecture”?

Claude Code · Claude Fable 5.2 · high, with Claude Max ($100) or Ultra ($200): Fable may use up to 50% of the weekly limit (on Pro, paid credits only). ≈ 2.9 agent-hours per 5-hour window · ≈ 14 per week on Claude Max 5x (central estimate). Or: Claude Code · Claude Opus 5.5 (default) · xhigh, with Claude Pro ($20), Max ($100) or Ultra ($200). ≈ 1.7 agent-hours per 5-hour window · ≈ 13 per week on Claude Pro (central estimate).

How long: Only for the plan: one session, until the plan or SPEC.md is written, reviewed and approved. Then: Stop there. Implement in a fresh session with Opus 5.5 at medium (default) — see Everyday coding — or hand the plan to a long run with /goal. Staying at xhigh to implement costs 2.5× more per task (AA index: $6.05 vs $2.45), and implementing with Fable 5.2 would spend its weekly cap on the part a cheaper model does well.

Why: A plan is one long reasoning pass, so effort pays off: Opus 5.5 scores 57.4 on the AA Intelligence Index (v5) at xhigh, above Claude Sonnet 5.5 at max (54.5 for $10.02) and just below its own max (59.1), for 40% less than max ($6.05 vs $10.04 per task). It works on Claude Pro.

Sources: Artificial Analysis model leaderboard · Snorkel LibraryDesignBench 1.0.0 · CWE-bench v2 · LiveBench · Details on the page

Which agent, model and effort for “Everyday coding”?

Claude Code · Claude Opus 5.5 (default) · medium (default), with Claude Pro ($20), Max ($100) or Ultra ($200). ≈ 2.3 agent-hours per 5-hour window · ≈ 17.6 per week on Claude Pro (central estimate).

How long: All day: medium, Claude Code's default, is the setting to leave on. One session per task (/clear between tasks), and end a session after two failed corrections. Then: A task that resists (a subtle change, two failed attempts): /effort high for that task only, or ultrathink for one turn, then back to medium. A bug you cannot explain: Debugging. A change you cannot describe in one sentence: plan it first (Advanced planning). Many small mechanical edits: High volume.

Why: The best quality for the quota on routine work. Benchmarks are made of hard tasks: there, high adds about 4 points over medium (CursorBench 4.0 52.5% → 56.0%, Terminal-Bench 4.0 (AA) 52.5% → 56.6%, FrontierCode 2.0 30.1% → 34.5%) for about 35% more quota, so about a quarter fewer agent-hours. On everyday tasks that medium already solves, that gain does not show; keep high for the tasks that resist.

Sources: Artificial Analysis model leaderboard · FrontierCode leaderboard data · Arena Code: WebDev leaderboard · Design Arena: Code overall · Details on the page

Which agent, model and effort for “Debugging”?

Claude Code · Claude Opus 5.5 (default) · high, with Claude Pro ($20), Max ($100) or Ultra ($200). ≈ 1.8 agent-hours per 5-hour window · ≈ 14 per week on Claude Pro (central estimate).

How long: Until the cause is found and the fix is proven by a test, then /effort medium again: high costs about 27% more per task than medium, too much to leave on all day. Then: Still stuck: add ultrathink to the next prompt (one turn at xhigh). Stuck after two attempts: /clear, then Claude Fable 5.2 at xhigh (Max or Ultra) with what you learned. Once fixed: back to Everyday coding.

Why: Diagnosis is a hard task, where effort shows. Terminal-Bench is mostly diagnosis in a terminal (failing builds, broken services, data to recover); on the Artificial Analysis run, Opus 5.5 gains 4 points from medium to high (52.5% → 56.6%) for $4.04 → $5.12 per task, about 27% more quota.

Sources: Artificial Analysis model leaderboard · CursorBench 4.0 · Vals AI benchmarks · Terminal-Bench 4.1 leaderboard · Details on the page

Which agent, model and effort for “Front-end and web UI”?

Claude Code · Claude Opus 5.5 (default) · medium (default), with Claude Pro ($20), Max ($100) or Ultra ($200). ≈ 2.3 agent-hours per 5-hour window · ≈ 17.6 per week on Claude Pro (central estimate).

How long: For pages, components and tweaks: medium, the default, already leads Design Arena, whose Opus 5.5 entries run at a default-like effort. Then: For the one page that must stand out (a landing page, a showcase): /effort max while you build it, the setting behind the Arena WebDev lead, then back to medium. Max burns about 35% more quota per hour (effort factor 1 vs 0.74).

Why: Top of Arena WebDev (1814.3 ±11.7, blind human votes, cutoff 2026-11-10), shared with Claude Fable 5.2 at xhigh (1835.8 ±22.4 on only 994 votes, ranks 1–2), and clearly ahead of Claude Sonnet 5.5 xhigh (1772.5), Gemini 4 Carbon high (1768.2, pre-release), GPT-6 Astra max (1764.4) and GPT-6.1 Sol max (1743.0). First on Arena Image-to-WebDev too (1751.9 ±19.2), level with Claude Sonnet 5.5 (1744.1).

Sources: Arena Code: WebDev leaderboard · Arena Code: Image-to-WebDev · Design Arena: Code overall · Design Arena: Website · Details on the page

Which agent, model and effort for “Long autonomous agent runs”?

Claude Code · Claude Opus 5.5 (default) · xhigh, with Claude Max ($100) or Ultra ($200) for long sessions. One agent never hits the 5-hour limit (≈ 8.6 agent-hours per window) · ≈ 82.2 agent-hours per week on Claude Max 5x (central estimate). Or: Codex · GPT-6.1 Sol (default) · xhigh, with ChatGPT Plus ($20) or Pro. ≈ 3.7 agent-hours per 5-hour window · ≈ 26.3 per week on ChatGPT Plus (Codex) (central estimate).

How long: For the whole run, until the /goal condition is met — hours if needed. Check the milestones instead of watching every step. Then: When the goal is met, end the run, review in a fresh context (/code-review), then go back to medium (Everyday coding) for follow-ups. If the run fails on the same error twice, stop it and re-plan instead of letting it burn quota.

Why: On the AA Coding Agent Index, which runs each model in its vendor's agent, Opus 5.5 at max in Claude Code is first (61.7 for $11.42 per task), ahead of Gemini 4 Carbon (59.4) and Claude Sonnet 5.5 (59.3); GPT-6.1 Sol at max in Codex scores 54.1 for $1.50, lower but for about an eighth of the cost. On the official tbench.ai leaderboard (Terminal-Bench 4.1, the 2026-11-09 regrade of 4.0), Opus 5.5 at max is first (65.76%, $14.30 per trial) and GPT-6.1 Sol at max (60.91%) edges GPT-6 Astra (60.30%) for a fifth of its cost per trial ($1.92 vs $9.90).

Sources: Artificial Analysis model leaderboard · Terminal-Bench 4.1 leaderboard · Artificial Analysis Coding Agent Index · Introducing Claude Sonnet 5.5 · Details on the page

Which agent, model and effort for “High volume, subagents, quick edits”?

Codex · GPT-6.1 Terra · medium (default), with ChatGPT Plus ($20). One agent never hits the 5-hour limit (≈ 6.6 agent-hours per window) · ≈ 47.4 agent-hours per week on ChatGPT Plus (Codex) (central estimate). Or: Claude Code · Claude Haiku 5.5 · medium (default), with Claude Pro ($20), Max ($100) or Ultra ($200). One agent never hits the 5-hour limit (≈ 16.1 agent-hours per window) · ≈ 122.3 agent-hours per week on Claude Pro (central estimate).

How long: As long as the tasks stay mechanical and fully specified: renames, types, docstrings, tests from a clear pattern. Test the prompt on 2–3 files, then fan out. Then: As soon as a task needs judgment (a design choice, an unclear bug), go back to Everyday coding. Only when the Terra quota runs out on trivial edits, GPT-6 Luna at max costs a little less, but scores far lower (Terminal-Bench 4.0 (AA) 12.6% against 32.3%). Never Claude Fable 5.2 for bulk work or subagents: it would empty its 50% weekly cap on edits a small model does as well.

Why: Quality first, then price, for easy tasks: bulk edits are mechanical and fully specified, so the default effort is enough. At medium, GPT-6.1 Terra scores 32.3% on Terminal-Bench 4.0 (AA) for $0.28 per task, 38.2 on the AA index for $0.17 and 42.0 on the AA Coding Agent Index for $0.39: ahead of Claude Haiku 5.5 at medium (15.2%, 24.6) and far ahead of GPT-6 Luna at max (12.6%, 27.9, 31.2).

Sources: Introducing GPT-6 Sol and Luna · Introducing GPT-6.1 Sol · Introducing GPT-6.1 Terra · Codex pricing · Details on the page

Which agent, model and effort for “Algorithms, proofs and competitive programming”?

Codex · GPT-6 Astra · max, with ChatGPT Pro ($100 / $200 / $500). ≈ 130.7 agent-hours per week on ChatGPT Pro 200 (Codex) (no 5-hour window; central estimate). Or: Codex · GPT-6.1 Sol (default) · max, with ChatGPT Plus ($20) or Pro. ≈ 3.7 agent-hours per 5-hour window · ≈ 26.3 per week on ChatGPT Plus (Codex) (central estimate).

How long: Only for the hard core — the algorithm, the proof, the optimization: one focused session. Then: Once the core is correct and tested, go back to Everyday coding for the surrounding code, I/O and tests.

Why: Quality first: GPT-6 Astra at max leads ALE-Bench (2951, against 2147 for Opus 5.5 at high), scores 100% on Vals IOI (Opus 5.5 95.1% ±4.9) and leads WeirdML v3 (42.2 at xhigh, Opus 5.5 31.2).

Sources: Vals AI benchmarks · ALE-Bench (Epoch AI Benchmarking Hub) · WeirdML v3 (Epoch AI Benchmarking Hub) · LiveBench · Details on the page

Which agent, model and effort for “Scientific and numerical code”?

Claude Code · Claude Fable 5.2 · high, with Claude Max ($100) or Ultra ($200): Fable may use up to 50% of the weekly limit (on Pro, paid credits only). ≈ 2.9 agent-hours per 5-hour window · ≈ 14 per week on Claude Max 5x (central estimate). Or: Claude Code · Claude Opus 5.5 (default) · xhigh, with Claude Max ($100) or Ultra ($200) for long sessions. One agent never hits the 5-hour limit (≈ 8.6 agent-hours per window) · ≈ 82.2 agent-hours per week on Claude Max 5x (central estimate).

How long: For the numerical core: the model, the formulas, the kernel, until it matches a reference result. Then: Once the core matches its reference, go back to Everyday coding (Opus 5.5 at medium) for I/O, plots and packaging: Fable 5.2 would spend its weekly cap on code a cheaper model writes as well.

Why: Quality first: Claude Fable 5.2 leads Terminal-Bench Science (AA) at high (70.0%; 73.3% at max), well ahead of Gemini 4 Carbon (64.8%), GPT-6 Astra (63.3% at max) and Opus 5.5 at xhigh (61.9%). High rather than xhigh (67.6%) or max, which costs far more for 3 points.

Sources: Artificial Analysis evaluation details · Artificial Analysis: Terminal-Bench-Science · Terminal-Bench-Science 0.1 (Snorkel mirror of the official board) · Artificial Analysis model leaderboard · Details on the page

Which agent, model and effort for “SVG illustrations and icons as code”?

Codex · GPT-6 Astra · xhigh, with ChatGPT Plus ($20, 5–45 Astra messages per 5 hours) or Pro ($100 / $200 / $500). ≈ 130.7 agent-hours per week on ChatGPT Pro 200 (Codex) (no 5-hour window; central estimate). Or: Claude Code · Claude Opus 5.5 (default) · high, with Claude Pro ($20), Max ($100) or Ultra ($200). ≈ 1.8 agent-hours per 5-hour window · ≈ 14 per week on Claude Pro (central estimate).

How long: While you create the assets — an icon set, an illustration, a logo — iterating with renders until it looks right. Then: For small edits to an existing SVG (colors, sizes), your everyday agent is enough.

Why: First general-purpose model in the Design Arena SVG arena: Elo 1421 ±9 over 1,808 blind battles (68.8% won), about 90 points above Gemini 4 Argon (1329), Kimi K3.1 Pro (1326) and Claude Opus 5.5 (1321 ±10, 1,261 battles): about seven standard errors, the only large-sample signal for SVG illustrations. Only the SVG-specialized Arrow 2 Telos, in no subscription, scores higher.

Sources: Design Arena: SVG · MacBook SVG Benchmark: 34 AI Models, Ranked by the Laptop They Draw · Managing usage with GPT-6 Astra in Work and Codex · Codex pricing · Details on the page

Which agent, model and effort for “Charts, dashboards and diagrams as code”?

Claude Code · Claude Opus 5.5 (default) · medium (default), with Claude Pro ($20), Max ($100) or Ultra ($200). ≈ 2.3 agent-hours per 5-hour window · ≈ 17.6 per week on Claude Pro (central estimate).

How long: While you build the chart or the first version of the diagram, iterating with renders until it reads right. Then: For later edits (a label, a new node, a color), your everyday agent is enough: a Mermaid diagram is ordinary text once it exists.

Why: Claude Opus 5.5 leads Design Arena data visualization (1390 ±14 over 700 blind battles), ahead of Claude Sonnet 5.5 (1362, on only 160 battles), Kimi K3 (1348), Claude Fable 5.1 (1342), Gemini 4 Argon (1333) and GPT-6 Astra (1325, 10th): about three standard errors ahead of Astra, the leader for SVG illustrations.

Sources: Design Arena: data visualization · MacBook SVG Benchmark: 34 AI Models, Ranked by the Laptop They Draw · Prompting Claude Opus 5.5 · Antigravity models · Details on the page

Which agent, model and effort for “3D, games and creative coding as code”?

Claude Code · Claude Opus 5.5 (default) · medium (default), with Claude Pro ($20), Max ($100) or Ultra ($200). ≈ 2.3 agent-hours per 5-hour window · ≈ 17.6 per week on Claude Pro (central estimate). Or: Codex · GPT-6 Astra · xhigh, with ChatGPT Pro ($100 / $200 / $500). ≈ 130.7 agent-hours per week on ChatGPT Pro 200 (Codex) (no 5-hour window; central estimate).

How long: While you build the scene and its interactions, at the default effort, medium. Then: For tuning values, colors and small fixes, /effort medium (default). Textures and icons as SVG: see SVG illustrations and icons.

Why: Games: the best Claude model on Design Arena game dev (1429, 2,666 battles), behind GPT-6 Astra at max (1466) and xhigh (1457), and ahead of Claude Fable 5.2 (1404), on browser games written as code.

Sources: Design Arena: Game dev · Arena Code: WebDev leaderboard · Design Arena: Code overall · Managing usage with GPT-6 Astra in Work and Codex · Details on the page

Is max effort worth it?

Rarely. The agent's default effort for everyday work; high for debugging, xhigh for planning and multi-hour runs; max only where you have measured a gain.

Opus 5.5: on Terminal-Bench 4.0 (AA) max equals xhigh (59.6%) for 49% more. Claude Fable 5.2: xhigh beats max on FrontierCode 2.0 (49.1% vs 45.7%) for a third of the cost. GPT-6 Astra: xhigh matches or beats max on Terminal-Bench 4.0 and DeepSWE (59.6% vs 59.1%; 74.1% vs 73.2%) for about 30% less. Anthropic warns that max can overthink.

Sources: FrontierCode leaderboard data · Artificial Analysis model leaderboard · DeepSWE v1.1 leaderboard · Model configuration · Details on the page

How much do coding agent subscriptions cost?

Monthly prices in US dollars, by agent:

  • Claude Code: Claude Pro $20, Claude Max 5x $100, Claude Ultra $200
  • Codex: ChatGPT Plus (Codex) $20, ChatGPT Pro 100 (Codex) $100, ChatGPT Pro 200 (Codex) $200, ChatGPT Pro 500 (Codex) $500
  • Antigravity: Google AI Pro (Antigravity) $19.99, Google AI Ultra 5x (Antigravity) $99.99, Google AI Ultra 20x (Antigravity) $199.99
  • Grok Build: X Subscription Premium (Grok Build) $30
  • Claude Code / ZCode: GLM Coding Plan Lite $18, GLM Coding Plan Pro $80, GLM Coding Plan Max $168
  • Qwen Code: Qwen Token Plan Lite $6, Qwen Token Plan Essential $10, Qwen Token Plan Standard $18, Qwen Token Plan Pro $68
  • MiMo Code: MiMo Token Plan Lite $6, MiMo Token Plan Standard $16, MiMo Token Plan Pro $50, MiMo Token Plan Max $100
  • GitHub Copilot: $10 / $39 / $100
  • OpenCode Go: $10 / $40 / $20 per member
  • Cursor: $8 / $30 / $100 / $200

Sources: Claude pricing · What is the Max plan? · What is the Ultra plan? · Codex pricing · About ChatGPT Pro tiers · Antigravity plans · Changes to Antigravity plans · Grok FAQ · Details on the page

How do the 5-hour and weekly limits work?

A rolling limit counted over 5 hours from the first request, then reset. Claude and ChatGPT (Codex) plans combine it with a weekly limit.

A second cap counted over 7 days, which stops heavy users even when each 5-hour window still has room. On Claude Max and Ultra plans, the Claude Fable models (5.2, 5.1, 5) may use at most 50% of it together. Claude Ultra has only this weekly limit (no 5-hour window once the account is verified).

The simulator estimates how many agent-hours each subscription gives, and when the quota runs out.

Sources: How do usage and length limits work? · Codex pricing · Claude Fable models on your plan · Details on the page

How many hours of agent work does a subscription give?

Central estimates of the simulator for the current picks (one agent working continuously):

  • Advanced planning and architecture (Claude Code · Claude Fable 5.2 · high): ≈ 2.9 agent-hours per 5-hour window · ≈ 14 per week on Claude Max 5x (central estimate).
  • Everyday coding (Claude Code · Claude Opus 5.5 (default) · medium (default)): ≈ 2.3 agent-hours per 5-hour window · ≈ 17.6 per week on Claude Pro (central estimate).
  • Debugging (Claude Code · Claude Opus 5.5 (default) · high): ≈ 1.8 agent-hours per 5-hour window · ≈ 14 per week on Claude Pro (central estimate).
  • Front-end and web UI (Claude Code · Claude Opus 5.5 (default) · medium (default)): ≈ 2.3 agent-hours per 5-hour window · ≈ 17.6 per week on Claude Pro (central estimate).
  • Long autonomous agent runs (Claude Code · Claude Opus 5.5 (default) · xhigh): One agent never hits the 5-hour limit (≈ 8.6 agent-hours per window) · ≈ 82.2 agent-hours per week on Claude Max 5x (central estimate).
  • High volume, subagents, quick edits (Codex · GPT-6.1 Terra · medium (default)): One agent never hits the 5-hour limit (≈ 6.6 agent-hours per window) · ≈ 47.4 agent-hours per week on ChatGPT Plus (Codex) (central estimate).
  • Algorithms, proofs and competitive programming (Codex · GPT-6 Astra · max): ≈ 130.7 agent-hours per week on ChatGPT Pro 200 (Codex) (no 5-hour window; central estimate).
  • Scientific and numerical code (Claude Code · Claude Fable 5.2 · high): ≈ 2.9 agent-hours per 5-hour window · ≈ 14 per week on Claude Max 5x (central estimate).
  • SVG illustrations and icons as code (Codex · GPT-6 Astra · xhigh): ≈ 130.7 agent-hours per week on ChatGPT Pro 200 (Codex) (no 5-hour window; central estimate).
  • Charts, dashboards and diagrams as code (Claude Code · Claude Opus 5.5 (default) · medium (default)): ≈ 2.3 agent-hours per 5-hour window · ≈ 17.6 per week on Claude Pro (central estimate).
  • 3D, games and creative coding as code (Claude Code · Claude Opus 5.5 (default) · medium (default)): ≈ 2.3 agent-hours per 5-hour window · ≈ 17.6 per week on Claude Pro (central estimate).

Sources: How do usage and length limits work? · Codex pricing · Details on the page

Which model leads the coding leaderboards right now?

The first place of each leaderboard, with the date of its last update:

Sources: Arena Code: WebDev leaderboard · Artificial Analysis Coding Agent Index · SWE-Bench Pro (public) V2 · LiveBench · Arena Agent Arena: Code · FrontierSWE V2 · SWE-Marathon v1.1 · Design Arena: Code overall · Design Arena: agentic web dev · Design Arena: Website · Arena Code: Image-to-WebDev · WeirdML v3 (Epoch AI Benchmarking Hub) · GSO (software optimization) · SWE Atlas (codebase Q&A, test writing, refactoring) · Snorkel Agentic Coding 2.0 · Snorkel SWE-bench CLI · Vals AI: ProgramBench · Vals AI: SRE-bench · CWE-bench v2 · Endor Labs Agent Security League · Artificial Analysis: Terminal-Bench-Science · Design Arena: UI components · Design Arena: Game dev · Design Arena: Mobile apps (agents) · Design Arena: SVG · Design Arena: data visualization · Best AI for coding · SWE-rebench · KiloBench · Terminal-Bench-Science 0.1 (Snorkel mirror of the official board) · Snorkel LibraryDesignBench 1.0.0 · Details on the page

Can I run a coding model locally?

Yes, on a powerful enough machine; the best choices by hardware, with their AA index and Terminal-Bench 4.0 (TB4) scores:

  • 24 GB GPU (RTX 4090 / 3090): Qwen3.8-27B Q4_K_M, 16.5 GB, 38–46 tok/s, AA 33.7 · TB4 5.6%
  • 32 GB GPU (RTX 5090): Qwen3.8-27B Q6 / NVFP4, 22–25 GB, 23–75 tok/s · MTP 98–134, AA 33.7 · TB4 5.6%
  • 128 GB unified memory (Strix Halo, DGX Spark, M5 Max): Qwen3.8-Flash-Next IQ4_XS, 93.7 GB, 11–37 tok/s · MTP 47–84, AA 39.8 · TB4 25.3%
  • 256 GB unified memory (M5 Ultra, M3 Ultra): GLM-5.3-Flash MLX 4-bit, 204 GB, 26–41 tok/s, AA 41.8 · TB4 32.8%

Sources: Qwen3.8-27B hardware tests · Qwen3.8-27B quantizations benchmarked · Artificial Analysis model leaderboard · Mac Studio M5 Ultra review · Details on the page

share

Share Coding Agent

Found it useful? Share the link in one click, or post the carousel that sums it up: every share credits the tool and links to it.

Share the link

Opens the network with the text and the link filled in: you only confirm. The text is also copied, for the networks that drop it (Facebook, LinkedIn on phones).

The carousel

The essentials of the page in slides, generated in your language, your mode and your color theme: change them at the top of the page and the slides follow. Every slide carries the name and the address of the tool.

Drawing the slides…

support

Support and feedback

A personal project, free and made with care: why this page exists, and how to reach me.

Why this page

It started with a question people keep asking me: “What should I use right now?” I test constantly myself, roughly every week — there is always something new, in open source and open weights, in paid models, in model updates and new models, for images as for code. I was tired of repeating the same advice to dozens of people, so I made this page: I keep it up to date, and you can come back to see what to use or not, what to do or avoid. Your reminder of the moment.

How the recommendations are ranked

The recommendations combine benchmark results and analyses with official provider guidance for each use case. Model and reasoning-effort choices take measured performance, cost and estimated quota consumption into account. Each choice links to dated sources, distinguishing independent measurements from vendor-published results.

Out of date? Tell me

This is a personal project. I usually update it as things happen, but on holiday or when other projects keep me busy, it can fall a little behind. If you notice it is out of date, send me a quick reminder — by email, WhatsApp, LinkedIn or any other way — and I will refresh everything right away. All my contacts are on my portfolio, at the root of this website.

Share your feedback

Share your own experience too, so this page gets better for everyone: send me your feedback, even critical — I love criticism, it is what makes things better.

Free, made with heart ❤️

Like most of my projects, this one is free and made with love. I hope it helped you; in any case, it was made with heart.

A word for the providers 🚜

If you are one of the providers listed in this tool, feel free to contact me to offer quota resets or credits: that is what I miss most. Each of us already holds 1 to 4 Max subscriptions. Nothing would make us happier than resets or credits. Get in touch: you will get a special mention here ❤️.

Contact me on josedacosta.net

references

References

Every figure on this page comes from these pages. For each one: what we took from it, and when.

  1. SWE-Marathon v1.1Abundant AIThird party

    pass@1 over 20 multi-hour tasks × 8 trials, mean cost per trial, tokens and errors for every agent × model × effort (data in the site's JS bundle, newest trial 2026-11-07; the deploy of 2026-10-15 hides 11 older models from the board, and so do we; Gemini 4 Carbon (Antigravity CLI) and Inkling 2 Small added in November 2026).

    Published Nov 7, 2026 · Accessed Nov 11, 2026

  2. Model Studio Coding PlanAlibaba CloudOfficial

    Coding Plan Pro $50: 6,000 requests per 5 hours, 45,000 per week, 90,000 per month; its model list has no Qwen3.8 model (up to qwen3.7-plus); sold in limited quantity, Lite discontinued.

    Published Sep 28, 2026 · Accessed Sep 28, 2026

  3. Token Plan (Personal Edition)Alibaba CloudOfficial

    Lite $6 / Essential $10 / Standard $18 / Pro $68 (list $8 / $16 / $25 / $80) for 11,500 / 25,500 / 45,000 / 180,000 credits per month; weekly quota removed on 2026-09-22, no 5-hour window; Qwen3.8 models cost 60% fewer credits from 22:00 to 08:00 (UTC+8).

    Published Sep 28, 2026 · Accessed Sep 28, 2026

  4. Download OpenCodeAnomaly (OpenCode)Official

    Surfaces: OpenCode Terminal (`opencode`), OpenCode Desktop (macOS, Windows, Linux), OpenCode Extensions (VS Code, Cursor, Zed, Windsurf, VSCodium).

    Accessed Sep 28, 2026

  5. OpenCode CLIAnomaly (OpenCode)Official

    `opencode run` for non-interactive use in scripts.

    Accessed Sep 29, 2026

  6. OpenCode GitHubAnomaly (OpenCode)Official

    GitHub integration: /opencode in issues and pull requests, run through GitHub Actions; GitLab is supported too.

    Accessed Sep 29, 2026

  7. OpenCode SDKAnomaly (OpenCode)Official

    OpenCode SDK (@opencode-ai/sdk) to control an OpenCode server from code.

    Accessed Sep 29, 2026

  8. OpenCode webAnomaly (OpenCode)Official

    `opencode web` starts a server (127.0.0.1 by default, network access protected by OPENCODE_SERVER_PASSWORD) and opens the interface in a browser; no hosted web surface.

    Accessed Sep 28, 2026

  9. Agent SDK overviewAnthropicOfficial

    Claude Agent SDK: the Claude Code agent as a library for your own programs.

    Accessed Sep 29, 2026

  10. Anthropic Economic Index: impact on software developmentAnthropicOfficial

    Languages (JS/TS 31%, HTML/CSS 28%) and task clusters (UI/UX components 12%, first) in 500,000 coding interactions.

    Published Apr 28, 2025 · Accessed Sep 28, 2026

  11. Claude Code /goalAnthropicOfficial

    `/goal <condition>`: a separate evaluator re-checks the condition after each turn.

    Accessed Sep 28, 2026

  12. Claude Code best practicesAnthropicOfficial

    Verification first (prompt, `/goal`, Stop hook, subagent); skip the plan for one-sentence changes; prune CLAUDE.md; `/clear` after two failed corrections; `/batch` fan-out; adversarial review.

    Accessed Sep 28, 2026

  13. Claude Code changelogAnthropicOfficial

    2.1.299 (2026-10-14): idle compaction, a 200K+ token conversation left idle about 54 minutes is compacted before its one-hour prompt cache expires; minimum set by CLAUDE_CODE_IDLE_COMPACT_MIN_TOKENS, off with "idleCompaction": false. 2.1.302 (2026-10-16): Idle compaction row in /config. 2.1.304 (2026-10-19): warning on Pro and Max when resuming an inactive conversation would use a noticeable share of the 5-hour limit. 2.1.311 (2026-10-26): ultrathink runs its turn at xhigh effort on Opus 5.5, Sonnet 5.5, Haiku 5.5 and Fable 5.1 in the interactive terminal. 2.1.312 (2026-10-27): usage messages say "5-hour limit" instead of "session limit". 2.1.316 (2026-10-31): CLAUDE_CODE_DISABLE_ULTRATHINK, CLAUDE_CODE_SUBAGENT_EFFORT (effort of subagents when nothing else sets one), ultrathink at xhigh in VS Code; Opus 4.8 fast mode deprecated, removed 2026-11-30. 2.1.317 (2026-11-03): subagents on another model use that model's saved or default effort, not the session's. 2.1.321 (2026-11-09): Claude Fable 5.2 ($8 / $40, $0.20 cache reads, 1M context) becomes the fable alias; saved Fable 5.1 picks move to it on Pro, Max and Ultra; Ultra named alongside Pro and Max. 2.1.322 (2026-11-10): Fable 5.2 fast mode on Vertex (global, us and eu regions).

    Published Nov 10, 2026 · Accessed Nov 11, 2026

  14. Claude Code commandsAnthropicOfficial

    `/code-review` (alias `/review`), `/simplify` (cleanup, not bug finding), `/batch`, `/goal`, `/effort`, `/doctor`, `/btw`, `/rewind`; workflow command syntax for `/diff`, `/init`, `/plan`, `/compact`, `/resume`, `/mcp` and `/hooks`.

    Accessed Sep 29, 2026

  15. Claude Code costsAnthropicOfficial

    Subagents and agent teams draw on the same usage limits; agent teams use about 7× more tokens when teammates run in plan mode.

    Accessed Sep 28, 2026

  16. Claude Code for VS CodeAnthropicOfficial

    Extension name "Claude Code for VS Code", id anthropic.claude-code, version 2.1.283.

    Published Sep 25, 2026 · Accessed Sep 28, 2026

  17. Claude Code GitHub ActionsAnthropicOfficial

    Surface name "Claude Code GitHub Actions" (anthropics/claude-code-action, @claude in issues and pull requests); "Claude Code GitLab CI/CD" for GitLab.

    Accessed Sep 29, 2026

  18. Claude Code in ChromeAnthropicOfficial

    `claude --chrome` lets the agent drive the browser; requires a subscription login (`/login`), unavailable with an API key.

    Accessed Sep 28, 2026

  19. Claude Code in SlackAnthropicOfficial

    Claude Code runs coding tasks started from Slack.

    Accessed Sep 29, 2026

  20. Claude Code on mobileAnthropicOfficial

    Start, monitor and steer Claude Code tasks from the Claude app for iOS and Android.

    Accessed Sep 29, 2026

  21. Claude Code permission modesAnthropicOfficial

    Auto mode is the starting mode since v2.1.283; `Shift+Tab` cycles auto → default → acceptEdits → plan; `/plan`.

    Accessed Sep 28, 2026

  22. Claude Code reviewAnthropicOfficial

    Local /code-review checks correctness; /review alias since 2.1.223. /simplify applies cleanup; --fix and --comment are explicit modification/publication options, not default review steps.

    Accessed Sep 29, 2026

  23. Claude Code workflowsAnthropicOfficial

    `ultracode` keyword or “use a workflow”; quota cost and the warning beyond 25 agents or 1.5M tokens.

    Accessed Sep 28, 2026

  24. Claude Code worktreesAnthropicOfficial

    `claude --worktree <name>` / `-w` for isolated parallel sessions.

    Accessed Sep 28, 2026

  25. Claude Code: run and verify your appAnthropicOfficial

    Bundled /run and /verify launch and exercise the app. /run-skill-generator records launch instructions; /verify may save a project skill. Availability depends on enabled bundled skills and project overrides.

    Accessed Sep 29, 2026

  26. Claude Fable 5.1 and Mythos 5.1AnthropicOfficial

    CursorBench 3.2.0 and Terminal-Bench 4.0 score and cost at each effort level for Fable 5.1, Fable 5 and Mythos 5.1.

    Published Sep 1, 2026 · Accessed Sep 28, 2026

  27. Claude Fable 5.2 and Claude Mythos 5.2 System CardAnthropicOfficial

    Coding scores at max effort unless stated: SWE-bench Pro 93.0%, Multilingual 96.6%, Multimodal 74.1%; DeepSWE v1.1 74.3%; Terminal-Bench 4.0 67.6% (Claude Code --bare, 660 trials, ±1.8; Opus 5.5 66.4% at xhigh, Fable 5.1 55.8%); FrontierCode 2.0 49.1% at xhigh and 45.7% at max (44.4% if 22 timed-out attempts count as zero); FrontierSWE v2 69.9%.

    Published Nov 9, 2026 · Accessed Nov 11, 2026

  28. Claude Fable 5.2 launch offersAnthropicOfficial

    From 2026-11-09: Pro and Team standard seats get every Fable model in their plan until 2026-11-16 (up to 50% of the weekly limit) plus one Fable reset; Max, Ultra and premium seats get one to three Fable resets by subscription length (under 6 months, 6–12, over 12), valid until 2027-03-08; a reset restores Fable usage on the weekly limit and refills the whole 5-hour window, without raising limits.

    Published Nov 9, 2026 · Accessed Nov 11, 2026

  29. Claude Fable 5.2 model pageAnthropicOfficial

    Model id claude-fable-5-2, released 2026-11-09, retirement not sooner than 2027-11-09; 1M context, 128K output; adaptive thinking always on; API default effort high; $8 input, $10 / $16 cache write (5 min / 1 h), $0.20 cache read, $40 output, batch 50%; knowledge cutoff June 2026; 30-day data retention required.

    Published Nov 9, 2026 · Accessed Nov 11, 2026

  30. Claude Fable models on your planAnthropicOfficial

    Fable 5.2, 5.1 and 5 on all paid plans; they use limits faster and are capped at 50% of the weekly limit on Max, Ultra and Team Premium; usage credits only on Pro and Team Standard; Fable 5.2 needs Claude Code 2.1.321 or later.

    Accessed Nov 11, 2026

  31. Claude Haiku 5.5 (Claude Platform docs)AnthropicOfficial

    Model id claude-haiku-5-5, 1M context, 128K output; $0.10 input, $0.01 cache read, $0.50 output per MTok up to 100K-token prompts ($0.50 / $0.05 / $2.50 above); efforts low to max, medium by default; adaptive thinking.

    Published Oct 7, 2026 · Accessed Oct 10, 2026

  32. Claude Haiku 5.5 System CardAnthropicOfficial

    Coding capabilities at max effort: SWE-bench Pro 64.8, Multilingual 83.7, Multimodal 30.7, FrontierCode Main 46.4 (45.8 at xhigh), Terminal-Bench 4.0 39.2, FrontierSWE v2 43.8; FrontierCode per effort in output tokens (figure 8.3.A).

    Published Oct 7, 2026 · Accessed Oct 10, 2026

  33. Claude Opus 5.5AnthropicOfficial

    Score and cost per task at each effort level for Opus 5.5, Fable 5.1 and Opus 5 on Terminal-Bench 4.0, FrontierCode v1.1 and CursorBench 4.0; five-hour limit change of 2026-09-22.

    Published Sep 22, 2026 · Accessed Sep 28, 2026

  34. Claude Opus 5.5 System CardAnthropicOfficial

    Headline coding scores (SWE-bench Pro, Terminal-Bench 4.0 at xhigh and max, DeepSWE v1.1, CursorBench 4.0) versus Opus 5, Fable 5.1 and GPT-6 Astra.

    Published Sep 22, 2026 · Accessed Sep 28, 2026

  35. Claude pricingAnthropicOfficial

    Prices of Free, Pro ($20, $17 a month billed yearly), Max ($100, 5× Pro), Ultra ($200, 10× Pro per week, no 5-hour limit for verified accounts), Team and Enterprise plans and what each includes; Fable on usage credits on Pro, up to half the weekly limit on Max and Ultra; monthly API credits on Max ($100), Ultra ($200) and Team (up to $500); API price table with Fable 5.2 $8 / $0.20 / $40.

    Accessed Nov 11, 2026

  36. Claude Sonnet 5 System CardAnthropicOfficial

    Headline coding scores of Sonnet 5 at max effort (SWE-bench Verified, SWE-bench Pro, Terminal-Bench 2.1).

    Published Jun 30, 2026 · Accessed Sep 28, 2026

  37. Claude Sonnet 5.5 (Claude Platform docs)AnthropicOfficial

    Model id claude-sonnet-5-5, 1M context, 128K output; $2 input, $0.10 cache read (cut from $0.20 on 2026-10-07), $10 output per MTok; efforts low to max, default high on the API and medium in Claude Code and the apps; thinking can no longer be disabled (between_tools instead).

    Published Sep 28, 2026 · Accessed Oct 10, 2026

  38. Claude Sonnet 5.5 System CardAnthropicOfficial

    Coding capabilities at max effort: DeepSWE v1.1 71.0, FrontierSWE v2 61.9.

    Published Sep 28, 2026 · Accessed Sep 29, 2026

  39. Continue local sessions from any device with Remote ControlAnthropicOfficial

    Remote Control continues a local Claude Code session from a phone, a tablet or any browser, through claude.ai/code or the Claude app.

    Accessed Sep 29, 2026

  40. EffortAnthropicOfficial

    Effort levels available per model and when to use each (xhigh for agentic coding over 30 minutes, low for subagents); Haiku 5.5 supports low to max, medium by default; Fable 5.2 supports low to max, high by default on the API (start at high, gains above it are uneven).

    Accessed Nov 11, 2026

  41. How AI is transforming work at AnthropicAnthropicOfficial

    Claude Code transcripts by task: feature implementation 36.9%, design and planning 9.9%; daily use for debugging 55%.

    Published Dec 2, 2025 · Accessed Sep 28, 2026

  42. How do usage and length limits work?AnthropicOfficial

    Five-hour session and weekly limits; no fixed message count; what drives consumption (model, effort, length, tools).

    Accessed Oct 10, 2026

  43. Introducing Claude Fable 5.2 and Claude Mythos 5.2AnthropicOfficial

    Score and cost per task at each effort level for Fable 5.2, Opus 5.5 and Fable 5.1 on FrontierCode 2.0 (also GPT-6 Astra and Gemini 4 Carbon; Fable 5.2 49.1% at xhigh for $14.19, 45.6% at max for $44.70), WANDR, BenchCAD 2.1, Humanity's Last Exam and Chartography; benchmark table (FrontierSWE v2 69.9%, Terminal-Bench-Science 0.1 74.6%); $8 / $40, fast mode $16 / $80; medium is the default in the Claude apps and Claude Code; Fable up to 50% of the weekly limit for Pro and Team standard seats during the launch week.

    Published Nov 9, 2026 · Accessed Nov 11, 2026

  44. Introducing Claude Haiku 4.5AnthropicOfficial

    SWE-bench Verified score of Haiku 4.5.

    Published Oct 15, 2025 · Accessed Sep 28, 2026

  45. Introducing Claude Haiku 5.5AnthropicOfficial

    Score and cost at each effort level for Haiku 5.5, Haiku 4.5, Sonnet 5.5 and GPT-6 Luna on OSWorld 2.1, GDPval-AA v2.1, Humanity's Last Exam and Terminal-Bench 4.0 (Sonnet 5.5 at $0.10 and $0.20 cache reads); headline Terminal-Bench 4.0 39.2 and FrontierCode 1.1 Main 46.4 at max; prices by prompt length; Sonnet 5.5 cache reads cut to $0.10; monthly API credits for Max and Team.

    Published Oct 7, 2026 · Accessed Oct 10, 2026

  46. Introducing Claude Sonnet 5AnthropicOfficial

    Release date, positioning and headline coding scores of Sonnet 5.

    Published Jun 30, 2026 · Accessed Sep 28, 2026

  47. Introducing Claude Sonnet 5.5AnthropicOfficial

    Score and cost at each effort level for Sonnet 5.5, Opus 5.5 and Sonnet 5 on Terminal-Bench 4.0 ($ per attempt), FrontierCode v1.1 and CursorBench 4.0 ($ per task); headline SWE-bench Pro 81.3 and Terminal-Bench 4.0 70.6 at max.

    Published Sep 28, 2026 · Accessed Sep 29, 2026

  48. Model configurationAnthropicOfficial

    Default model (Opus 5.5 on every plan since v2.1.280; no Fable model is a default); default effort per model (medium on Fable 5.2, Opus 5.5, Sonnet 5.5 and Haiku 5.5); fable alias → Fable 5.2 and saved Fable 5.1 picks moved to it on Pro, Max and Ultra (v2.1.321); haiku alias → Haiku 5.5 since v2.1.293; ultrathink raises one terminal turn to xhigh (v2.1.311+, Fable 5.2 included).

    Accessed Nov 11, 2026

  49. Models overviewAnthropicOfficial

    Model IDs, release dates, context windows, max output and status of every current Claude model: current lineup Fable 5.2, Opus 5.5, Sonnet 5.5, Haiku 5.5; Fable 5.1, Fable 5 and Haiku 4.5 legacy, still available.

    Accessed Nov 11, 2026

  50. Models, usage, and limits in Claude CodeAnthropicOfficial

    Opus uses meaningfully more quota than Sonnet; higher effort reaches limits faster.

    Accessed Oct 10, 2026

  51. Monthly API credits for Max and Team plansAnthropicOfficial

    Max 5x $100, Max 20x $200, Team $20 per Standard and $100 per Premium seat pooled up to $500 a month; Claude Platform only (API, Managed Agents, Agent SDK with an API key); not interactive Claude Code nor extra usage; plan limits unchanged.

    Published Oct 7, 2026 · Accessed Oct 10, 2026

  52. Optimizing for cost and intelligenceAnthropicOfficial

    Cost per solved task on a 478-problem SWE-bench Pro subset for Opus 5.5, Fable 5.1 and Sonnet 5 configurations; effort trade-offs for Opus 5.5; Haiku 5.5: GPQA Diamond 85% at about a twentieth of Opus 5.5's cost per question.

    Accessed Oct 10, 2026

  53. Platforms and integrationsAnthropicOfficial

    Claude Code surfaces: CLI, Desktop, VS Code, JetBrains, Web (claude.ai/code) and Mobile; Slack and Chrome are integrations, not surfaces.

    Accessed Sep 28, 2026

  54. PricingAnthropicOfficial

    Input, cache write, cache read and output prices per million tokens; batch and fast-mode prices (Fable 5.2 fast mode $16 / $80 since 2026-11-09; Opus 5.5 $8 / $40, out of research preview since 2026-10-15). Fable 5.2 $8 / $0.20 / $40, Fable 5.1 $10 / $0.25 / $50 (cache reads 0.025×); Sonnet 5.5 cache reads $0.10 (0.05×) since 2026-10-07; Haiku 5.5 priced by prompt length (over 100K tokens).

    Accessed Nov 11, 2026

  55. Prompting Claude Opus 5.5AnthropicOfficial

    Start at medium; reserve xhigh and max for work with a measured gain; name the finish line; name the frontend styles to avoid.

    Accessed Sep 28, 2026

  56. Run Claude Code programmaticallyAnthropicOfficial

    Non-interactive use of Claude Code in scripts and pipelines (`claude -p`).

    Accessed Sep 29, 2026

  57. Use Claude Code DesktopAnthropicOfficial

    Claude Code Desktop is the Code tab of the Claude desktop app (Chat, Cowork, Code) on macOS, Windows and Linux (beta).

    Accessed Sep 28, 2026

  58. Use Claude Code in the cloudAnthropicOfficial

    Browser surface at claude.ai/code, "also called Claude Code on the web"; cloud sessions on Pro, Max, Team and Enterprise.

    Accessed Sep 28, 2026

  59. What is the Max plan?AnthropicOfficial

    Max at $100: five times Pro usage every five hours and every week, a separate weekly limit for Fable; monthly billing only; $100 a month of API credits; Max 20x replaced by the Ultra plan.

    Accessed Nov 11, 2026

  60. What is the Ultra plan?AnthropicOfficial

    The Max 20x plan is now called Ultra, existing subscribers moved automatically with no price change; $200 a month, monthly only; 10 times Pro usage every week; no five-hour limit for verified accounts (a temporary limit until verification); a weekly limit across all models and a separate weekly limit for Fable; $200 a month of API credits.

    Accessed Nov 11, 2026

  61. 5-hour session limits increaseAnthropic (ClaudeDevs on X)Official

    Five-hour limits up 20% on 2026-09-22; Opus 5.5 goes 25% further within limits because it is priced lower.

    Published Sep 22, 2026 · Accessed Sep 28, 2026

  62. Arena Agent Arena: CodeArenaThird party

    Net improvement over an average model (with CI), sessions and the five signal scores of each model × effort on live coding-agent sessions (snapshot of 2026-11-08, 909,397 sessions, 57 models; flight payload; Bash Reliability replaced Bash Recovery on 2026-10-12).

    Published Nov 8, 2026 · Accessed Nov 11, 2026

  63. Arena Code: Image-to-WebDevArenaThird party

    Rating, 95% confidence bounds and votes of every model building a web page from an image (vote cutoff 2026-11-10, 210,832 votes, 66 models; entries[] in the payload).

    Published Nov 10, 2026 · Accessed Nov 11, 2026

  64. Arena Code: WebDev leaderboardArena (LMArena)Third party

    Rating, 95% confidence interval, votes and rank of every model on blind pairwise web-development votes (entries[] in the server-rendered payload; vote cutoff 2026-11-10, 995,980 votes, 157 models). The page now sits behind a Cloudflare challenge: load it in a headed browser.

    Published Nov 10, 2026 · Accessed Nov 11, 2026

  65. Announcing Artificial Analysis Intelligence Index v5Artificial AnalysisThird party

    Nine components of Index v5 (HLE-Diamond replaces Humanity's Last Exam, FrontierCode 2.0 and Terminal-Bench-Science added), 60% on private test sets.

    Published Oct 27, 2026 · Accessed Nov 11, 2026

  66. Artificial Analysis Coding Agent IndexArtificial AnalysisThird party

    Coding Agent Index per harness × model (Claude Code, Codex, Grok Build, Kimi Code, Antigravity, OpenCode, Muse Code, Devin Fusion, GitHub Copilot, Mistral Vibe, Mirror, DeepSeek Harness): score on FrontierCode v1.1, Terminal-Bench 4.0 and SWE-Atlas-QnA-AA, cost, tokens and time per task (35 rows; reward-hacking audit materialized 2026-11-10).

    Published Nov 10, 2026 · Accessed Nov 11, 2026

  67. Artificial Analysis evaluation detailsArtificial AnalysisThird party

    Per-model sub-scores embedded in the page payload: Humanity's Last Exam, CritPt, SciCode, GPQA and Terminal-Bench 4.0 for each model × effort.

    Accessed Nov 11, 2026

  68. Artificial Analysis model leaderboardArtificial AnalysisThird party

    Intelligence Index v5.0.1 and AA's own Terminal-Bench 4.0 run for every model × effort, with cost per task and output tokens per task (JSON embedded in the page; 734 models, 119 with a v5 score, newest release 2026-11-09: Claude Fable 5.2). Costs AA publishes as exactly 0 (no API price) are left empty.

    Accessed Nov 11, 2026

  69. Artificial Analysis: Terminal-Bench-ScienceArtificial AnalysisThird party

    pass@1 (mean of 3 repeats) of 135 model × effort rows from the embedded catalogue (terminalBenchScience); Claude rows marked "Default Fallback".

    Accessed Nov 11, 2026

  70. Claude Max weekly limit measurementsClassmethodThird party

    Weekly budget of Max 5x and Max 20x in API-equivalent dollars (two accounts, /usage vs ccusage).

    Published Aug 20, 2026 · Accessed Sep 28, 2026

  71. Codex usage researchcodexusage.devThird party

    Codex Plus 5-hour and weekly budgets in credits and dollars, cost per busy hour by effort level.

    Published Sep 22, 2026 · Accessed Sep 28, 2026

  72. Coding plan trackercodingplan.orgThird party

    Current Kimi plan names and CNY prices (Go, Plus, Pro, Max) and the mapping from legacy names.

    Published Sep 24, 2026 · Accessed Sep 28, 2026

  73. FrontierCode leaderboard dataCognitionThird party

    FrontierCode 2.0 (132 tasks, from 2026-10-27; data/frontier-code-2.0/data.json): score, pass rate, cost, tokens and flagged rate for every model × effort. FrontierCode v1 and v1.1 archive (frontiercode-leaderboard/data.json): score, pass rate and cost, main and extended subsets. Changelog: Gemini 4 Carbon 2026-10-28; SWE-2 Flash and Muse Spark Open 2026-11-04; Kimi K3.1 Pro 2026-11-05; Claude Fable 5.2 and Inkling 2 Small 2026-11-09.

    Published Nov 9, 2026 · Accessed Nov 11, 2026

  74. CWE-bench v2Collinear AIThird party

    Deterministic pass@1 and pass@4 with 95% CI and mean cost per rollout of 11 models in their agents, all at high reasoning, on the combined 250 tasks and on the public (100) and held-out (150) sets (v2, results as of 2026-11-02; table data attributes and data-v2.js). Archives: v1 at /v1 (120 tasks, programmatic and judge-panel pass@1/pass@4, updated 2026-10-28 with Gemini 4 Carbon) and v0 at /v0 (100 tasks).

    Published Nov 2, 2026 · Accessed Nov 11, 2026

  75. Agent SecurityCursorOfficial

    Tool approvals, MCP permissions and the distinction between approval classification and an isolation boundary.

    Accessed Sep 29, 2026

  76. BugbotCursorOfficial

    Surface name "Bugbot": Cursor's pull request review bot on GitHub.

    Accessed Sep 29, 2026

  77. Building a better BugbotCursorOfficial

    Bugbot reviews more than two million pull requests per month.

    Accessed Sep 29, 2026

  78. Cloud AgentsCursorOfficial

    Web surface name "Cloud Agents": agents that run in isolated cloud VMs, in parallel, without the local machine.

    Accessed Sep 28, 2026

  79. Cursor CLICursorOfficial

    Terminal surface name "Cursor CLI", started with `agent` (macOS, Linux, WSL, Windows).

    Accessed Sep 28, 2026

  80. Cursor for iOSCursorOfficial

    Cursor's cloud agents from an iPhone.

    Accessed Sep 29, 2026

  81. Cursor in JetBrains IDEsCursorOfficial

    Cursor's agent in IntelliJ IDEA, PyCharm, WebStorm and other JetBrains IDEs through the Agent Client Protocol (ACP); needs a paid Cursor plan.

    Accessed Sep 28, 2026

  82. Cursor in LinearCursorOfficial

    Cursor agents started from Linear issues.

    Accessed Sep 29, 2026

  83. Cursor in SlackCursorOfficial

    Cursor agents started from Slack.

    Accessed Sep 29, 2026

  84. Cursor pricingCursorOfficial

    The X Subscription (2026-10-12): Plus $8, Premium $30, Super $100, Ultra $200; one weekly pool for Cursor models shared with Grok (4×, 15×, 50× Plus); third-party models at API list price from prepaid credit on Premium and up.

    Published Oct 12, 2026 · Accessed Oct 17, 2026

  85. Cursor SDKCursorOfficial

    Cursor SDK for TypeScript and Python (@cursor/sdk).

    Accessed Sep 29, 2026

  86. CursorBench 4.0CursorThird party

    Score, cost per task, tokens and steps for every model × effort (server-rendered results table). Rows added without a changelog entry: Gemini 4 Argon and Muse Wonder (after 2026-10-10), DeepSeek V4.1 Flash 1017 (after 2026-10-17), Gemini 4 Carbon and Claude Fable 5.2 (November 2026); the changelog's latest entry is still 2026-10-07, so the published date is inferred.

    Published Nov 10, 2026 · Accessed Nov 11, 2026

  87. Legacy Cursor plansCursorOfficial

    Hobby, Start, Pro $20, Pro+ $60, Ultra $200 and Teams continue for existing customers with two monthly pools (Cursor Models, Other Models); comparable X Subscription tiers Premium, Super, Ultra.

    Accessed Oct 17, 2026

  88. Using Headless CLICursorOfficial

    Cursor CLI in scripts and automation, without the interactive interface.

    Accessed Sep 29, 2026

  89. What is The X Subscription?CursorOfficial

    One subscription for Grok, Cursor and X, launched on 2026-10-12: Grok and Cursor share one weekly included usage pool; third-party models in Cursor (Claude, GPT) are not included and are billed at API list price from the first use, paid from prepaid credit; switching is optional and permanent, and existing Cursor, SuperGrok and X Premium plans continue.

    Published Oct 12, 2026 · Accessed Oct 17, 2026

  90. DeepSWE v1.1 leaderboardDatacurveThird party

    pass@1 and mean cost per task for every model × effort (mini-swe-agent harness, 4 runs).

    Accessed Sep 28, 2026

  91. GLM-5.3 in Claude Code: real cost on the Coding PlanDécodeur IAThird party

    Lite plan in Claude Code: a 3-hour session used 70% of a 2,000-credit window; one endpoint with tests took 18 minutes and 210 credits.

    Published Aug 29, 2026 · Accessed Sep 28, 2026

  92. Design Arena: agentic web devDesign ArenaThird party

    Elo, win rate and battles for the full-stack and frontend agentic web-app arenas (full-stack from /api/leaderboard, frontend from the embedded board). No standard error published since 2026-10-10 (btStdErr null).

    Published Nov 11, 2026 · Accessed Nov 11, 2026

  93. Design Arena: Code overallDesign ArenaThird party

    Elo, battles, win rate and mean generation time of 195 models across every code category (category allcategories, 1,950,799 votes; no standard error published). Claude Opus 5.5 369 s, Claude Sonnet 5.5 390 s, Claude Fable 5.2 732 s, GPT-6.1 Sol 228 s, GPT-6.1 Terra 132 s, GPT-6 Astra xhigh 574 s, max 1,093 s, Gemini 4 Argon 205 s, Kimi K3.1 Pro 1,121 s.

    Published Nov 11, 2026 · Accessed Nov 11, 2026

  94. Design Arena: data visualizationDesign ArenaThird party

    Elo, battles and win rate of 156 models on blind human votes for chart and dashboard code (POST /api/leaderboard, category dataviz, 90,018 votes; no standard error published; models with fewer than 100 battles are no longer listed since 2026-10-16).

    Published Nov 11, 2026 · Accessed Nov 11, 2026

  95. Design Arena: Game devDesign ArenaThird party

    Elo, battles, win rate and mean generation time of 193 models on browser games (category gamedev, 318,370 votes; no standard error published). Claude Opus 5.5 483 s, Claude Sonnet 5.5 486 s, Claude Fable 5.2 919 s, GPT-6.1 Sol 228 s, GPT-6.1 Terra 123 s, GPT-6 Astra xhigh 603 s, max 1,133 s, Gemini 4 Argon 198 s, Kimi K3.1 Pro 1,076 s.

    Published Nov 11, 2026 · Accessed Nov 11, 2026

  96. Design Arena: Mobile apps (agents)Design ArenaThird party

    Elo, battles and win rate of 58 agents building mobile apps (arenaType agents, category mobileapps, 45,185 votes; no standard error published).

    Published Nov 11, 2026 · Accessed Nov 11, 2026

  97. Design Arena: SVGDesign ArenaThird party

    Elo, battles and win rate of 139 models on blind human votes for SVG generation (POST /api/leaderboard, category svg, 211,612 votes; standard error no longer published since 2026-10-10).

    Published Nov 11, 2026 · Accessed Nov 11, 2026

  98. Design Arena: UI componentsDesign ArenaThird party

    Elo, battles and win rate of 186 models on UI components (category uicomponent, 114,714 votes; no standard error published).

    Published Nov 11, 2026 · Accessed Nov 11, 2026

  99. Design Arena: WebsiteDesign ArenaThird party

    Elo, battles and win rate of 206 models on single-shot websites (POST /api/leaderboard, category website, 1,190,870 votes; no standard error published).

    Published Nov 11, 2026 · Accessed Nov 11, 2026

  100. Endor Labs Agent Security LeagueEndor LabsThird party

    Secure and functional pass rates of 34 harness × model rows (Claude Code, Codex, Cursor, Muse Code, SWE-Agent), extending the SusVibes benchmark (200 tasks, 77 CWE classes); newest row 2026-10-24 (Muse Code + Muse Wonder); board re-scored between 2026-10-16 and 2026-10-24.

    Published Oct 24, 2026 · Accessed Nov 11, 2026

  101. ALE-Bench (Epoch AI Benchmarking Hub)Epoch AIThird party

    ALE-Bench Performance score (algorithm optimization) and cost per model × effort (ale_bench_external.csv): GPT-6 Astra max 2951, GPT-6.1 Sol max 2744, GPT-6.1 Terra max 2609 at $0.41, GPT-6 Sol max 2462, Claude Opus 5.5 high 2147 (the only Opus 5.5 effort measured).

    Published Nov 11, 2026 · Accessed Nov 11, 2026

  102. WeirdML v3 (Epoch AI Benchmarking Hub)Epoch AIThird party

    WeirdML v3 agentic score and cost per run per model, agent and effort (weirdml_v3_external.csv; benchmark released 2026-09-18).

    Published Nov 11, 2026 · Accessed Nov 11, 2026

  103. Factory pricingFactoryOfficial

    Droid Pro, Plus and Max prices and per-model multipliers.

    Published Sep 28, 2026 · Accessed Sep 28, 2026

  104. Kimi plans in USDgeotoolbox.aiThird party

    International USD prices of Kimi plans and their Kimi Code usage multipliers (1×, 5×, 15×, 30×).

    Published Sep 5, 2026 · Accessed Sep 28, 2026

  105. 60 million Copilot code reviews and countingGitHubOfficial

    Copilot code review passed 60 million reviews; usage grew 10x since its April 2025 launch and accounts for more than one in five code reviews on GitHub.

    Published Mar 5, 2026 · Accessed Sep 29, 2026

  106. About GitHub Copilot CLIGitHubOfficial

    Terminal surface name "GitHub Copilot CLI", started with `copilot` (Linux, macOS, Windows); `copilot -p` runs it programmatically.

    Accessed Sep 28, 2026

  107. About GitHub Copilot cloud agentGitHubOfficial

    Copilot cloud agent works in the background from the agents panel on GitHub.com, issues or VS Code and opens pull requests; it can also be started from Microsoft Teams and Slack (public preview) and from Azure Boards, Jira or Linear.

    Accessed Sep 28, 2026

  108. About GitHub Copilot code reviewGitHubOfficial

    Surface name "Copilot code review": Copilot reviews pull requests, identifies issues and suggests fixes.

    Accessed Sep 29, 2026

  109. Asking GitHub Copilot questions in GitHubGitHubOfficial

    Web surface: Copilot Chat on any GitHub page and at github.com/copilot.

    Accessed Sep 29, 2026

  110. Asking GitHub Copilot questions in GitHub MobileGitHubOfficial

    Copilot in GitHub Mobile: coding questions, repositories and pull requests from a phone.

    Accessed Sep 29, 2026

  111. Claude Fable 5.2 in GitHub CopilotGitHubOfficial

    Claude Fable 5.2 generally available on 2026-11-09 for Copilot Pro+, Max, Business and Enterprise on the same surfaces; billed at provider list price; requires data retention by Anthropic (policy off by default for organizations; zero data retention for eligible enterprises until the end of 2026).

    Published Nov 9, 2026 · Accessed Nov 11, 2026

  112. Copilot SDKGitHubOfficial

    GitHub Copilot SDK to build Copilot-powered apps.

    Accessed Sep 29, 2026

  113. Gemini 4 Argon in GitHub CopilotGitHubOfficial

    Gemini 4 Argon generally available on 2026-10-13 for Copilot Pro+, Max, Business and Enterprise in VS Code, Visual Studio, Copilot CLI, cloud agent, Copilot app, JetBrains, Xcode and Eclipse; billed at provider list price.

    Published Oct 13, 2026 · Accessed Oct 17, 2026

  114. Gemini 4 Carbon in GitHub CopilotGitHubOfficial

    Gemini 4 Carbon generally available on 2026-11-10 for Copilot Pro+, Max, Business and Enterprise in VS Code, Visual Studio, Copilot CLI, coding agent, Copilot app, github.com, GitHub Mobile, JetBrains, Xcode and Eclipse; billed at provider list price; on by default for organizations unless disabled.

    Published Nov 10, 2026 · Accessed Nov 11, 2026

  115. GitHub Action dependentsGitHubThird party

    Repositories using each official action on 2026-10-11: anthropics/claude-code-action 24,081; JetBrains/junie-github-action 47.

    Accessed Oct 11, 2026

  116. GitHub Copilot plansGitHubOfficial

    Copilot Pro, Pro+, Max, Business and Enterprise prices and the AI Credits included with each (unchanged: $10 / $39 / $100 with $15 / $70 / $200); Gemini 4 Carbon, Gemini 4 Argon and Claude Fable 5, 5.1 and 5.2 on Pro+ and Max, not on Free or Pro.

    Published Sep 28, 2026 · Accessed Nov 11, 2026

  117. Installing the GitHub Copilot extension in your environmentGitHubOfficial

    IDE surface name "GitHub Copilot extension" (VS Code, Visual Studio, JetBrains, Xcode, Vim/Neovim, Azure Data Studio).

    Accessed Sep 28, 2026

  118. Models and pricing for GitHub CopilotGitHubOfficial

    Per-token AI Credit prices per model (1 credit = $0.01); Gemini 4 Carbon and Gemini 4 Argon $2 / $0.10 / $10 per MTok, promotional until 2026-12-01 (Carbon then $4 / $0.20 / $20); Claude Fable 5.2 $8 / $0.20 / $40 (cache write $10).

    Accessed Nov 11, 2026

  119. Route and orchestrate models with GitHub Copilot FusionGitHubOfficial

    Copilot Fusion generally available on 2026-10-28 on all Copilot plans (VS Code, Copilot app, CLI, JetBrains): Standard, High and Max efforts route and orchestrate enabled models; billed per model used at its standard rate, 30% off for a limited time on paid plans; Fusion Max scored 60 on the AA Coding Agent Index at about half the cost per task of Claude Code with Opus 5.5 Max.

    Published Oct 28, 2026 · Accessed Nov 11, 2026

  120. Pro 20x weekly meter measurementGitHub user reportThird party

    Weekly meter moved from 64% to 95% for $609 of API-equivalent usage on Pro 20x.

    Published Sep 10, 2026 · Accessed Sep 28, 2026

  121. Three weekly meters on Pro 5xGitHub user reportThird party

    Three full weekly Pro 5x meters = $959.87 API-equivalent (about $320 per week) and 708 M tokens on GPT-5.6 Sol and Astra.

    Published Sep 7, 2026 · Accessed Sep 28, 2026

  122. Weekly allowance value on Pro 20x (GPT-5.6 Sol vs GPT-6 Astra)GitHub user reportThird party

    API-equivalent value of a full weekly Pro 20x allowance: about $1,925 on GPT-5.6 Sol, $1,191–1,234 on GPT-6 Astra.

    Published Sep 8, 2026 · Accessed Sep 28, 2026

  123. Claude Code usage meter measurementsGitHub user reportsThird party

    Controlled A/B measurements of API-equivalent dollars per 1% of the 5-hour and weekly meters (interactive vs headless), Pro and Max.

    Published Sep 27, 2026 · Accessed Sep 28, 2026

  124. Antigravity CLI best practicesGoogleOfficial

    Verification loops first; explore, plan, then execute; `/rewind`, `/fork`, parallel subagents.

    Accessed Sep 28, 2026

  125. Antigravity CLI installationGoogleOfficial

    Antigravity CLI binary `agy`, install script antigravity.google/cli/install.sh.

    Accessed Sep 28, 2026

  126. Antigravity IDE ExtensionsGoogleOfficial

    Extensions for VS Code (Marketplace "Google Antigravity", Google.google-antigravity), Visual Studio, JetBrains, Zed and Xcode.

    Accessed Sep 28, 2026

  127. Antigravity modelsGoogleOfficial

    Models per plan: Gemini 4 Carbon (Low/Medium/High, default High) on Ultra and Enterprise only; Gemini 4 Argon (same levels) on paid Google AI Pro (or a trial with eligible premium devices), Ultra and Enterprise; Claude Opus 5.5 and Sonnet 5.5 (thinking) on Pro (non-trial), Ultra and Enterprise; Gemini 3.8 Flash on every plan; Gemini 3.7 Flash, 3.6 Flash and 3.1 Pro on Enterprise only; Claude 4.6 models and GPT-OSS-120b to be removed on 2026-11-18 (was 2026-11-02). Two usage pools with five-hour and weekly remaining percentages.

    Accessed Nov 11, 2026

  128. Antigravity plansGoogleOfficial

    Quota wording per plan: weekly refresh on Free, five-hour refresh until a weekly limit on Pro and Ultra.

    Accessed Oct 17, 2026

  129. Antigravity Remote ControlGoogleOfficial

    Remote Control: a browser dashboard, installable on a phone with push notifications, that drives Antigravity 2.0 and CLI sessions running on your own machines; no hosted web agent.

    Accessed Sep 29, 2026

  130. Changes to Antigravity plansGoogleOfficial

    Ultra $100 = 5× and $200 = 20× the Pro token allowance; the Gemini pool is drawn down at API prices; Claude and GPT models use a separate fixed pool.

    Published May 19, 2026 · Accessed Oct 10, 2026

  131. Gemini 4 Argon in Google AntigravityGoogleOfficial

    Gemini 4 Argon rolls out in the Antigravity desktop app and CLI for Google AI Pro and Ultra subscribers and eligible Gemini Enterprise subscriptions, from 2026-10-13.

    Published Oct 13, 2026 · Accessed Oct 17, 2026

  132. Gemini 4 Argon: Built to solve your hardest problemsGoogleOfficial

    General availability on 2026-10-13: Gemini API and AI Studio, Gemini app and AI Mode for Google AI Pro and Ultra, Antigravity; introductory price $2 / $10 per MTok (cached input 95% off) until 2026-11-30, then $4 / $20; 1M unified input-output token limit.

    Published Oct 13, 2026 · Accessed Oct 17, 2026

  133. Gemini 4 Argon: our next era of frontier intelligenceGoogleOfficial

    Restricted rollout to cyber defenders (Fairwind Program) before paid API customers and Google AI Ultra; introductory price $2 input / $10 output per MTok, cached input 95% off, then $4 / $20; 1M-token output limit; DeepSWE v1.1 77.9%.

    Published Sep 30, 2026 · Accessed Oct 10, 2026

  134. Gemini 4 Carbon and /swarmGoogleOfficial

    Gemini 4 Carbon for Google AI Ultra subscribers in the Antigravity app, CLI and IDE extensions from 2026-11-10, and in Gemini for business; /swarm multi-agent teams in preview ("consume quota much faster than a standard chat").

    Published Nov 10, 2026 · Accessed Nov 11, 2026

  135. Gemini 4 Carbon is now availableGoogleOfficial

    General availability on 2026-11-10: Gemini API paid tier (AI Studio, Android Studio), Gemini for business; Google AI Ultra subscribers in Antigravity, the Gemini app and AI Mode (not Pro). Argon stays the choice for most work; Carbon uses more tokens. #1 on CursorBench 4.0, WANDR and AutomationBench; /swarm preview in Antigravity; intro price ends 2026-11-30.

    Published Nov 10, 2026 · Accessed Nov 11, 2026

  136. Gemini 4 Carbon: our next frontier modelGoogleOfficial

    Restricted rollout to a subset of cyber defenders (Fairwind Program) from 2026-10-28; same introductory price as Argon, $2 / $10 per MTok, cached input 95% off, then $4 / $20. Results table (same values as the evaluation PDF): Terminal-Bench 4.0 (AA) 65.2, FrontierSWE 2 61.2, SWE-sweep 9.6; WebDev Arena #4 (1768, +89); cheating on Terminal-Bench 4.0 reduced to 0.0%.

    Published Oct 28, 2026 · Accessed Nov 11, 2026

  137. Gemini API deprecationsGoogleOfficial

    gemini-3.7-flash deprecated on 2026-10-08 with requests routed to gemini-3.8-flash; gemini-3.5-flash routed to gemini-3.6-flash; gemini-2.5-pro routed to gemini-3.8-flash and gemini-2.5-flash to gemini-3.5-flash-lite from 2026-12-01 (announced 2026-11-02); gemini-4-carbon (2026-11-10) and gemini-4-argon (2026-10-13) with no shutdown date.

    Published Nov 10, 2026 · Accessed Nov 11, 2026

  138. Gemini API pricingGoogleOfficial

    Prices per million tokens: Gemini 4 Carbon and Gemini 4 Argon both $2 input / $0.10 cached / $10 output through 2026-11-30, then $4 / $0.20 / $20 (Batch and Flex half, Priority 1.8×, no long-context tier, not on the free tier; Contributor tier for Argon only); Gemini 3.8 Flash (introductory until 2026-12-31) and 3.1 Pro Preview. The page moved from ai.google.dev (301) and no longer prints a date; published = the Carbon entry of the API changelog.

    Published Nov 10, 2026 · Accessed Nov 11, 2026

  139. Gemini thinkingGoogleOfficial

    Thinking levels and defaults per Gemini model (Gemini 4 Carbon and Gemini 4 Argon: low, medium, high, default medium); Gemini 4 shares one 1,048,576-token window between input and output.

    Accessed Nov 11, 2026

  140. Google AI subscriptionsGoogleOfficial

    Prices of Google AI Plus, Pro, Ultra 5x and Ultra 20x (unchanged: $4.99, $19.99, $99.99, $199.99); Gemini 4 Argon included from AI Pro, Gemini 4 Carbon on Ultra only ("highest access" to both); Antigravity rate-limit wording per plan.

    Accessed Nov 11, 2026

  141. Google Antigravity docs: choose your surfaceGoogleOfficial

    Surfaces: Antigravity 2.0 (standalone desktop app), Antigravity CLI (TUI), Antigravity SDK and Antigravity IDE.

    Accessed Sep 28, 2026

  142. What's new in Gemini 4 Argon (Gemini API)GoogleOfficial

    Model id gemini-4-argon (stable); thinking levels low, medium (default, recommended for complex code and agentic use) and high; 1,048,576-token unified context, 65,536 output per synchronous request and up to 1M with long decode; implicit caching only (95% off); not on the free tier. Since Gemini 4 Carbon (2026-11-10): "efficient" model with fewer thinking tokens per turn than Carbon at the same price, and usable on the Google AI Pro plan in AI Studio (Carbon needs Ultra). The guide moved from ai.google.dev/gemini-api/docs/latest-model, now a model overview.

    Published Oct 13, 2026 · Accessed Nov 11, 2026

  143. What's new in Gemini 4 Carbon (Gemini API)GoogleOfficial

    Model id gemini-4-carbon (stable, released 2026-11-10, no shutdown date); thinking levels low, medium (default) and high; same 1,048,576-token unified window, 65,536 output per synchronous request and up to 1M with long decode, and same per-token price as Argon; spends more thinking tokens per turn than Argon; implicit caching only; AI Studio subscription access on AI Ultra only (Argon on AI Pro and Ultra).

    Published Nov 10, 2026 · Accessed Nov 11, 2026

  144. Fairwind ProgramGoogle DeepMindOfficial

    Vetted Fairwind partners (governments, critical infrastructure, core platforms) get Gemini 4 Argon without cyber guardrails, standalone or in CodeMender; they had exclusive access from 2026-09-30 until general availability on 2026-10-13.

    Accessed Oct 17, 2026

  145. Gemini 3.1 Pro model cardGoogle DeepMindOfficial

    SWE-bench Verified, SWE-bench Pro and Terminal-Bench 2.0 scores of Gemini 3.1 Pro (high thinking).

    Published Feb 19, 2026 · Accessed Sep 28, 2026

  146. Gemini 3.8 Flash evaluationGoogle DeepMindOfficial

    Coding benchmark table of Gemini 3.8 Flash versus Opus 5, Sonnet 5 and GPT-5.6, with the settings used.

    Published Sep 2, 2026 · Accessed Sep 28, 2026

  147. Gemini 4 Argon model cardGoogle DeepMindOfficial

    Sparse mixture-of-experts, 1M combined input-output tokens, knowledge cutoff March 2026; distributed in the Gemini app, Gemini API, AI Studio, Gemini for business, Antigravity and AI Mode. Its results table is the same image as the 2026-09-30 evaluation PDF (no new figure).

    Published Oct 13, 2026 · Accessed Oct 17, 2026

  148. Gemini 4 Argon model evaluationGoogle DeepMindOfficial

    Benchmark table of Gemini 4 Argon at its highest thinking setting versus GPT-6 Astra, Claude Fable 5.1 and Opus 5.5 (DeepSWE v1.1 77.9, Terminal-Bench 4.0 57.4, FrontierSWE v2 55.0, Vibe Code Bench 91.9), with the harness and settings per benchmark.

    Published Sep 30, 2026 · Accessed Oct 10, 2026

  149. Gemini 4 Carbon model cardGoogle DeepMindOfficial

    Based on Gemini 4 Argon; 1M combined input-output tokens; knowledge cutoff March 2026; uses more thinking tokens than Argon; Ultra in consumer surfaces and Antigravity, paid API tiers, Fairwind for verified defenders. Its results image is byte-identical to the evaluation PDF's.

    Published Nov 10, 2026 · Accessed Nov 11, 2026

  150. Gemini 4 Carbon model evaluationGoogle DeepMindOfficial

    Benchmark table of Gemini 4 Carbon and Argon at their highest thinking setting versus GPT-6 Astra, Claude Opus 5.5 and Fable 5.1: AA Index v5.0 59, Vals Index v2.2 70.4, SWE-sweep 9.6, FrontierSWE 2 61.2, Terminal-Bench 4.0 (AA) 65.2 (Argon 57.1), Agents' Last Exam v1.1 77.4, OSWorld-2.1 offline 81.5; harness and source per benchmark.

    Published Nov 10, 2026 · Accessed Nov 11, 2026

  151. GSO (software optimization)GSO team (UC Berkeley)Third party

    Hack-adjusted and raw Opt@1 scores per model × effort in OpenHands (leaderboard.json, last_updated 2026-09-27; rows after the 2026-09-27 elicitation change are not comparable with older ones).

    Published Sep 27, 2026 · Accessed Oct 11, 2026

  152. Qwen3.8-27B hardware testsHardware CornerThird party

    Local decode speeds of Qwen3.8-27B on RTX 4090, 3090, 5090 and M5 Max at several context lengths.

    Published Aug 17, 2026 · Accessed Sep 28, 2026

  153. Junie, the coding agent by JetBrainsJetBrainsOfficial

    Surfaces: Junie CLI (terminal) and Junie in JetBrains IDEs and Android Studio, plus GitHub and GitLab; plans AI Pro and AI Ultimate.

    Accessed Sep 28, 2026

  154. JetBrains Marketplace downloadsJetBrains MarketplaceThird party

    Downloads on 2026-10-11: GitHub Copilot (17718) 54,950,931; Junie (26104) 35,516,694; Claude Code [Beta] (27310) 4,885,789.

    Accessed Oct 11, 2026

  155. AI coding agent adoption 2026JetBrains ResearchThird party

    Share of professional developers using each AI coding tool at work, May–July 2026 (Developer Ecosystem Survey 2026, 15,000+ respondents, reweighted by region): Claude Code 39%, GitHub Copilot 21%, Codex 16%, Cursor 12%, JetBrains AI and/or Junie about 9%, OpenCode 7%, Antigravity 6%.

    Published Aug 15, 2026 · Accessed Sep 28, 2026

  156. KiloBenchKiloThird party

    Terminal-Bench 2.0 completion and cost per full attempt in the Kilo agent (embedded chart points).

    Published Sep 28, 2026 · Accessed Sep 28, 2026

  157. LiveBenchLiveBenchThird party

    LiveBench 2.0 releases of 2026-11-06 (shown as updated 2026-11-09), 2026-10-25 and 2026-10-14 (data/<release>/table.csv, categories.json, cost.csv, breakdown.json): suite scores, weighted overall (×2 long agentic, ×1.5 coding and design from 2026-10-25) and cost per task, each model at its highest supported effort (Claude Fable 5.2 at xhigh). Earlier releases were republished with changed costs and added models. LiveBench v1 release 2026-06-25 (under /v1/): coding and agentic-coding category scores and cost per question.

    Published Nov 9, 2026 · Accessed Nov 11, 2026

  158. Best AI for codingllm-stats (ZeroEval)Third party

    TrueSkill coding rating (μ − 3σ) over 68 benchmark games for every model (initialIndexData in the payload; 237 models reviewed). The CDN can serve stale copies even with a cache-busting query: check dateModified of each fetch.

    Published Nov 11, 2026 · Accessed Nov 11, 2026

  159. Mac Studio M5 Ultra reviewMacStoriesThird party

    Local decode speed of GLM-5.3-Flash on the M5 Ultra.

    Published Sep 21, 2026 · Accessed Sep 28, 2026

  160. Muse CodeMetaOfficial

    Muse Code plans (Everyday, High, Power) and prompts per 5 hours.

    Published Sep 28, 2026 · Accessed Sep 28, 2026

  161. MiniMax M PlanMiniMaxOfficial

    M Plan replaced Token Plan on 2026-09-30 at the same prices: Go $22, Explore $55, Build $132 a month (minimax.io/m-plan); Explore 3× and Build 7.5× Go usage; 5-hour and weekly windows, no sizes; text model M3.1 Flash Preview.

    Published Sep 30, 2026 · Accessed Oct 11, 2026

  162. MiniMax Token PlanMiniMaxOfficial

    Token Plan prices and 5-hour / weekly windows. Renamed M Plan on 2026-09-30 (legacy subscribers keep Token Plan).

    Published Sep 28, 2026 · Accessed Oct 10, 2026

  163. Introducing Mistral Large 4Mistral AIOfficial

    Public preview of Mistral Large 4 (1T parameters, 52B active, open weights by the end of October); coding scores from the text: DeepSWE v1.1 61.7%, SWE-Atlas QnA 59.4%, Terminal-Bench 4.0 28.3%, Coding Agent Index 49.8% (runs by Artificial Analysis, effort not stated). Charts are images with rounded labels.

    Published Oct 6, 2026 · Accessed Oct 10, 2026

  164. Mistral Large 4 (Le Chonk): Built for Coders, Trusted by DefendersMistral AIOfficial

    Open weights of the improved Mistral Large 4 ("Le Chonk") on 2026-10-29, three reasoning modes (low replaces none, medium, high by default), $0.70 / $0.07 cached / $2.10 per MTok. From the embedded chart JSON: Terminal-Bench 4.0 41.4% (28% for the preview), Artificial Analysis Coding Agent Index 48 at high (8.5M tokens per task) and 35 at medium (5.8M), Cognition's FrontierCode 2.0 19.7% (the high-effort run in Cognition's data, https://cognition.com/data/frontier-code-2.0/data.json: low 9.5, medium 8.8, Mistral Vibe harness), AA Cyber Index 50.9%.

    Published Oct 29, 2026 · Accessed Nov 11, 2026

  165. Mistral Large 4 (Mistral docs)Mistral AIOfficial

    Ids mistral-large-4 / mistral-large-4-0, version 26.10, GA, release date 2026-11-05, 1.05T total and 52B active parameters, 1M context, open weights (Mistral Ecosystem Open Weight License); $0.70 input, $0.07 cached, $2.10 output per MTok (preview list price $1.36 / $0.14 / $4.18). Reasoning guide: reasoning_effort low, medium or high (high recommended); none, minimal, xhigh and max refused. Mistral Medium 3.5 deprecated 2026-10-31 (retirement 2027-04-30), replaced by Mistral Large 4.

    Published Nov 5, 2026 · Accessed Nov 11, 2026

  166. Mistral pricingMistral AIOfficial

    Pro $14.99 with $25.50 of API credits a month (was $30 on 2026-09-28; seen with France preselected in the country selector), Education $5.99 with $15, Team $24.99 per user ($50 minimum); fair usage, PAYG past the limit at API rates.

    Published Sep 28, 2026 · Accessed Nov 11, 2026

  167. Vibe Code Web: limits and lifecycleMistral AIOfficial

    Free 2 and paid 100 Vibe Code Web sessions a day, 24-hour sessions; Vibe Code Web and the CLI with a Mistral model use Mistral Medium 3.5 (unchanged on 2026-11-11).

    Accessed Nov 11, 2026

  168. Kimi Agent SDKMoonshot AIOfficial

    Kimi Agent SDK: a programmatic interface to Kimi Code CLI.

    Accessed Sep 29, 2026

  169. Kimi API pricingMoonshot AIOfficial

    Input, cached input, output and cache-write prices per million tokens for Kimi K3.1 Pro ($2 / $0.20 / $10), K3.1 Pro Highspeed ($4 / $0.20 / $20), Kimi K3 ($3 / $0.30 / $15) and K2.7 Code; 1,048,576-token context for the K3 series.

    Accessed Oct 27, 2026

  170. Kimi Code Desktop changelogMoonshot AIOfficial

    Kimi Code Desktop 1.0.0 released on 2026-09-17; 1.0.3 on 2026-09-23 (macOS, Windows).

    Published Sep 23, 2026 · Accessed Sep 28, 2026

  171. Kimi Code membershipMoonshot AIOfficial

    Five-hour rolling window, monthly cap shared across Kimi services, extra usage top-ups; K3.1 Pro and K3 from Plus, 1M context and Highspeed from Pro.

    Accessed Oct 27, 2026

  172. Kimi Code modelsMoonshot AIOfficial

    Models in Kimi Code (K3.1 Pro, K3.1 Pro Highspeed at 2× quota, K3, K3 256K, K2.8 Preview, K2.7 Code Highspeed at 3× quota), effort levels with their defaults, and plan access.

    Accessed Oct 27, 2026

  173. Kimi Code overviewMoonshot AIOfficial

    Official clients Kimi Code Desktop, Kimi Code CLI (`kimi`) and Kimi Code for VS Code; JetBrains and Zed through ACP (`kimi acp`).

    Accessed Sep 28, 2026

  174. Kimi Code: What's NewMoonshot AIOfficial

    K3.1 Pro series live in Kimi Code on 2026-10-23: K3.1 Pro on Plus / Moderato and above (1M context from Pro / Allegretto), K3.1 Pro Highspeed on Pro / Allegretto and above; low / high / max with high as the default.

    Published Oct 23, 2026 · Accessed Oct 27, 2026

  175. Kimi K3Moonshot AIOfficial

    Coding benchmark table of Kimi K3 (max effort) versus Fable 5, GPT-5.6 Sol, Opus 4.8 and GLM-5.2, with harnesses used.

    Published Jul 16, 2026 · Accessed Sep 28, 2026

  176. Kimi K3.1 Pro (Hugging Face)Moonshot AIOfficial

    Open weights of Kimi K3.1 Pro (repository created 2026-11-02, Kimi K3.1 Pro License), MXFP4 weights with quantization-aware training, and the launch benchmark table (Terminal-Bench 4.0 in the Claude Code harness).

    Published Nov 2, 2026 · Accessed Nov 11, 2026

  177. Kimi K3.1 Pro quickstartMoonshot AIOfficial

    Model ids kimi-k3.1-pro and kimi-k3.1-pro-highspeed, 1M context, thinking always on, reasoning_effort low / high / max with max as the API default, access after a $1 top-up.

    Accessed Oct 27, 2026

  178. Kimi K3.1 Pro Tech Blog: Advancing Open IntelligenceMoonshot AIOfficial

    Benchmark table of Kimi K3.1 Pro at max effort in the Claude Code harness (Terminal-Bench 4.0 65.7, DeepSWE v1.1 75.5, SWE-Marathon v1.1 51.7, Terminal-Bench-Science 0.1 38.1) against Fable 5.1, GPT-6 Astra, Opus 5.5, DeepSeek-V4.1-Flash-1017 and GLM-5.3; Terminal-Bench 4.0 score and cost per task at high (52.8%, $4.16) and max (65.7%, $5.25); K3.1 Pro Highspeed 2-3× faster; weights by 2026-11-04.

    Published Oct 24, 2026 · Accessed Oct 27, 2026

  179. Kimi membership pricingMoonshot AIOfficial

    CNY monthly prices of the membership tiers (legacy names).

    Accessed Sep 28, 2026

  180. AIDev-pop: pull requests opened by coding agentsMSR 2026 (arXiv)Third party

    Types of 33,596 agent pull requests: features 43.0%, fixes 24.1%, docs 11.6%, tests 7.0%, refactors 6.8%.

    Published Feb 3, 2026 · Accessed Sep 28, 2026

  181. SWE-rebenchNebiusThird party

    Resolved rate, pass@5 and cost per task on the 2026-05-15 → 2026-07-01 window (embedded items).

    Published Jul 24, 2026 · Accessed Sep 28, 2026

  182. npm downloads API: coding agent CLIs, week of 2026-10-04npmThird party

    Weekly downloads, 2026-10-04 to 2026-10-10: @openai/codex 23,140,547; @anthropic-ai/claude-code 12,482,566; opencode-ai 2,958,079; @github/copilot 1,562,659; @qwen-code/qwen-code 68,764; @xai-official/grok 60,608; @moonshot-ai/kimi-code 47,815; @minimax-ai/code 11,370; @qoder-ai/qodercli 6,861; @mimo-ai/cli 2,107 (same endpoint for each package).

    Published Oct 10, 2026 · Accessed Oct 11, 2026

  183. npm downloads API: coding agent SDKs, week of 2026-10-04npmThird party

    Weekly downloads, 2026-10-04 to 2026-10-10: @opencode-ai/sdk 38,413,827; @anthropic-ai/claude-agent-sdk 12,106,186; @openai/codex-sdk 2,145,052; @github/copilot-sdk 1,475,033; @cursor/sdk 893,901; @qwen-code/sdk 3,564; @moonshot-ai/kimi-agent-sdk 1,737.

    Published Oct 10, 2026 · Accessed Oct 11, 2026

  184. Cross-provider quota calibrationoh-my-pi (GitHub)Third party

    Tokens and API-equivalent dollars per 1% of the Anthropic, OpenAI and Antigravity meters, measured on one machine.

    Published Sep 26, 2026 · Accessed Sep 28, 2026

  185. A Practical Approach to Verifying Code at ScaleOpenAIOfficial

    Codex code review handled more than 100k external pull requests per day as of October 2025; authors changed code after 52.7% of its comments.

    Accessed Sep 29, 2026

  186. About ChatGPT Pro tiersOpenAIOfficial

    Pro 100, Pro 200 and Pro 500 ($500, Ultrafast). Updated 2026-11-03: Pro 100 and Pro 500 include twice their previous Work and Codex allowance; the Pro 200 reduction announced on 2026-09-29 is canceled, every Pro 200 subscription has the higher allowance (the lower one doubled to match); Plus unchanged; no five-hour limit on Pro.

    Published Nov 3, 2026 · Accessed Nov 11, 2026

  187. API credits for ChatGPT subscribersOpenAIOfficial

    Since 2026-10-28, monthly API credits for eligible subscriptions: Pro 100 $100, Pro 200 $200, Pro 500 $500, Business Standard $20 and Premium $100 per paid seat (capped at $500 per workspace); unused credits expire each cycle; Codex usage through API keys is not covered and the credits do not raise ChatGPT or Codex limits.

    Published Oct 28, 2026 · Accessed Nov 11, 2026

  188. API pricingOpenAIOfficial

    Standard, batch, flex, fast and ultrafast prices per million tokens for GPT-6 Astra, GPT-6.1 Sol, GPT-6.1 Terra, GPT-6 Sol and Luna; Ultrafast is 6× Standard (Astra and 6.1 Sol only).

    Accessed Oct 27, 2026

  189. ChatGPT rate card (credit-based pricing)OpenAIOfficial

    Credits per million tokens for each model (25× the API price; GPT-6.1 Sol 50 / 2.5 / 250); Fast 2× and Ultrafast 6× when paid with credits; 2.5× and 8× of included usage.

    Accessed Oct 10, 2026

  190. Codex best practicesOpenAIOfficial

    Goal / Context / Constraints / Done when; start GPT-6.1 Sol at the client's default effort, Luna High, Astra Light; `/plan`, interview and PLANS.md; update AGENTS.md after a repeated mistake; worktrees for parallel tasks (the .md copy is gone: read the HTML page).

    Accessed Oct 10, 2026

  191. Codex changelog: Codex joins the ChatGPT desktop appOpenAIOfficial

    Codex app 26.707: Codex became part of the ChatGPT desktop app on macOS and Windows; the former "Codex app" name is retired.

    Published Jul 9, 2026 · Accessed Sep 28, 2026

  192. Codex CLIOpenAIOfficial

    Name "Codex CLI", command `codex`, installer chatgpt.com/codex/install.sh (npm @openai/codex).

    Accessed Sep 28, 2026

  193. Codex cloudOpenAIOfficial

    Web surface name "Codex cloud", opened at chatgpt.com/codex (GitHub, GitLab, Linear, Slack).

    Accessed Sep 28, 2026

  194. Codex code reviewOpenAIOfficial

    Review in CLI and desktop, local changes or a base branch; reviewer reports findings without changing the working tree. GitLab merge requests in preview and automatic cloud reviews.

    Accessed Oct 10, 2026

  195. Codex developer commandsOpenAIOfficial

    CLI syntax for /init, /plan, /diff, /mention, /compact, /resume, /fork, /mcp and /skills; codex review --uncommitted includes staged, unstaged and untracked files. Review target flags and a custom prompt are mutually exclusive.

    Accessed Sep 29, 2026

  196. Codex glossaryOpenAIOfficial

    Surface names: ChatGPT desktop app, Codex CLI, Codex IDE extension (VS Code, JetBrains, Cursor, Windsurf) and Codex cloud (web).

    Accessed Sep 28, 2026

  197. Codex IDE extensionOpenAIOfficial

    Extension id openai.chatgpt for VS Code, Cursor and Windsurf; JetBrains through AI Chat; Xcode through its coding assistant.

    Accessed Sep 28, 2026

  198. Codex modelsOpenAIOfficial

    GPT-6.1 Sol recommended (Plus to Enterprise, not Free/Go), effort Light to Ultra; starting presets: client default for 6.1 Sol, Luna High, Astra Light; GPT-6.1 Terra for writing, visual understanding and computer use (Codex CLI 0.159.0 or later, desktop app, Work on the web, API); surfaces per model.

    Accessed Oct 27, 2026

  199. Codex pricingOpenAIOfficial

    Plan prices, estimated local messages per 5 hours on Plus and Standard Business (GPT-6.1 Sol 15–160, GPT-6.1 Terra 35–285), no 5-hour limit on Pro; credit rates per model (GPT-6.1 Terra 25 / 1.25 / 125); Fast 2.5× and Ultrafast 8× of included usage.

    Accessed Oct 27, 2026

  200. Codex promptingOpenAIOfficial

    Task context, expected and actual behaviour, constraints and verification for coding-agent requests. Used to write editorial workflow templates, not universal commands.

    Accessed Sep 29, 2026

  201. Codex RemoteOpenAIOfficial

    Start, guide and review Codex tasks from the ChatGPT mobile app while Codex runs on a connected Mac or Windows PC.

    Accessed Sep 29, 2026

  202. Codex rust-v0.161.0OpenAIOfficial

    GPT-6.1 Sol becomes the default model of the bundled Codex catalog.

    Published Oct 7, 2026 · Accessed Oct 10, 2026

  203. Codex SDKOpenAIOfficial

    Codex SDK to drive Codex from your own programs.

    Accessed Sep 29, 2026

  204. Codex speed (Fast and Ultrafast)OpenAIOfficial

    Fast uses included usage at 2.5× (credits 2×); Ultrafast for GPT-6 Astra and GPT-6.1 Sol on Pro $500 and eligible Enterprise/Edu at 8× (credits 6×).

    Accessed Oct 10, 2026

  205. DevDay 2026 RecapOpenAIOfficial

    Pro 500 plan at 25× the Plus allowance with Ultrafast; Ultrafast up to 8× faster token generation (300 tokens per second) in Codex, 6× in the API; GPT-6.1 Sol launch; Codex cloud, CLI and Code Review updates.

    Published Sep 29, 2026 · Accessed Oct 10, 2026

  206. GPT-6 AstraOpenAIOfficial

    Terminal-Bench 4.0 score, cost, output tokens and latency at each effort for GPT-6 Astra, GPT-5.6 Sol and Fable 5.1; FrontierCode 1.1 Extended; output tokens per DeepSWE task.

    Published Sep 3, 2026 · Accessed Sep 28, 2026

  207. GPT-6 Astra model page (and /gpt-6.1-sol, /gpt-6.1-terra, /gpt-6-sol, /gpt-6-luna)OpenAIOfficial

    Release dates, context windows, knowledge cutoffs and supported effort values.

    Accessed Oct 10, 2026

  208. GPT-6.1 Sol model pageOpenAIOfficial

    Model id gpt-6.1-sol; 1,050,000 context, 128,000 output, cutoff 2026-04-30; efforts low to max (no none), default medium; $2 input, $0.10 cached, $2.50 cache write, $10 output; Fast 2×, Ultrafast 6×, batch and flex 50%.

    Published Sep 29, 2026 · Accessed Oct 10, 2026

  209. GPT-6.1 Terra model pageOpenAIOfficial

    Model id gpt-6.1-terra; 1,050,000 context (922,000 input), 128,000 output, cutoff 2026-06-01; efforts none to max, default medium; $1 input, $0.05 cached, $1.25 cache write (30 min), $5 output; above 272K input tokens 2× input and 1.5× output; Fast 2×, batch and flex 50%.

    Published Oct 26, 2026 · Accessed Oct 27, 2026

  210. Introducing GPT-6 Sol and LunaOpenAIOfficial

    DeepSWE v1.1 and FrontierCode 1.1 Main score and cost per task at each effort for GPT-6 Astra, Sol, Luna, GPT-5.6, Opus 5 and Fable 5.

    Published Sep 22, 2026 · Accessed Oct 10, 2026

  211. Introducing GPT-6.1 SolOpenAIOfficial

    Score and cost per task at each effort for GPT-6.1 Sol, GPT-6 Sol and GPT-6 Astra on DeepSWE v1.1 (6.1 Sol 75.22 at high), Terminal-Bench Science 0.1, OSWorld 2.0, AutomationBench, GDP.pdf and factual errors; prices $2 / $0.10 / $10.

    Published Sep 29, 2026 · Accessed Oct 10, 2026

  212. Introducing GPT-6.1 TerraOpenAIOfficial

    Score and cost per task at each effort for GPT-6.1 Terra, GPT-6.1 Sol, GPT-6 Astra and others on Terminal-Bench 4.0 (Terra 50.3 at max, $1.84), Terminal-Bench-Science 0.1, Agents' Last Exam v1, OSWorld 2.1, MMMU-Pro and factual errors; prices $1 / $0.05 / $5; Plus, Pro, Enterprise and Edu in Work and Codex.

    Published Oct 26, 2026 · Accessed Oct 27, 2026

  213. Managing billing and seats in ChatGPT BusinessOpenAIOfficial

    Business Standard and Premium seat prices, annual and monthly.

    Accessed Sep 28, 2026

  214. Managing usage with GPT-6 Astra in Work and CodexOpenAIOfficial

    Five-hour and weekly windows, shared allowance across Codex and Work, effort advice (Astra Low can beat Sol High); no 5-hour limit on Pro 100, 200 and 500; Ultrafast on Pro 500 at 8× included usage.

    Accessed Oct 10, 2026

  215. Non-interactive modeOpenAIOfficial

    `codex exec` runs Codex without the interactive interface, in scripts and CI.

    Accessed Sep 29, 2026

  216. Review GitHub pull requests with CodexOpenAIOfficial

    Codex code review on GitHub pull requests (@codex review, automatic reviews); a Codex GitHub Action also exists.

    Accessed Sep 29, 2026

  217. Use Codex in LinearOpenAIOfficial

    Codex tasks started from Linear issues.

    Accessed Sep 29, 2026

  218. Use Codex in SlackOpenAIOfficial

    Codex tasks started from Slack.

    Accessed Sep 29, 2026

  219. OpenCode GoOpenCodeOfficial

    Go ($10), Go Plus ($40) and Go Teams ($20 per member, $5 a month more usage than Go on every model) prices, requests per 5 hours, week and month and monthly dollar limit per model (open-weight models plus Grok 4.7, GPT-6 Luna, Haiku 5.5 and Gemini 4 Argon at half price until 2026-11-30); DeepSeek V4.1 Flash 43,330 / 86,670 / 45,500 per 5 h. 5-hour limit 20% and weekly 50% of the monthly limit.

    Published Sep 28, 2026 · Accessed Nov 11, 2026

  220. OpenRouter rankings: top models by taskOpenRouterThird party

    Share of spend and tokens per task over 30 days (code: general implementation 10.1%, debugging 5.2%, file edits 4.3%, review 2.8%, front-end UI 2.0%; math 0.8%), from /api/frontend/v1/rankings/task-spend.

    Published Sep 27, 2026 · Accessed Sep 28, 2026

  221. MacBook SVG Benchmark: 34 AI Models, Ranked by the Laptop They DrawPlaycodeThird party

    Quality order judged by eye over 249 runs (34 models, first answer kept per model × effort), updated 2026-11-09: 1st Claude Opus 5.5 (best drawing at max on the Batch API, 255,957 tokens, $2.56; streamed max returns an empty file; xhigh $1.72, high $0.40), 2nd Claude Fable 5.2 (best at medium $3.91, high $3.05; xhigh and max streamed drew nothing, $5.12 each), 3rd GPT-6 Astra (max $2.29, every effort assembles; xhigh $0.98), 4th Claude Sonnet 5.5, 5th GPT-6.1 Sol (xhigh $0.21), 6th GPT-6 Sol (max $0.39), 7th Claude Fable 5.1, 8th GPT-6.1 Terra, 9th Gemini 4 Argon, 10th Kimi K3.1 Pro.

    Published Nov 9, 2026 · Accessed Nov 11, 2026

  222. FrontierSWE V2Proximal LabsThird party

    Mean, best and worst of 5 trials, cost per trial and duration on 34 long-horizon tasks.

    Published Nov 9, 2026 · Accessed Nov 11, 2026

  223. Qwen3.8-27B quantizations benchmarkedQuesmaThird party

    Quality of Qwen3.8-27B by quantization level on Terminal-Bench 2.1 (Q4_K_M matches BF16, Q2 loses about 10 points).

    Published Aug 26, 2026 · Accessed Sep 28, 2026

  224. Headless ModeQwen (Alibaba)Official

    Qwen Code in scripts and automation without the interactive interface.

    Accessed Sep 29, 2026

  225. Qwen CodeQwen (Alibaba)Official

    Surfaces (CLI, desktop app, VS Code / Zed / JetBrains, web UI), release v0.24.6, 28,185 GitHub stars.

    Published Sep 26, 2026 · Accessed Sep 28, 2026

  226. Qwen Code ChannelsQwen (Alibaba)Official

    Qwen Code Channels: control the agent from Telegram, WeChat, QQ, DingTalk, WeCom and Feishu (Lark).

    Accessed Sep 29, 2026

  227. Qwen Code Desktop v0.24.6Qwen (Alibaba)Official

    Desktop app name "Qwen Code Desktop" (macOS, Windows, Linux).

    Published Sep 26, 2026 · Accessed Sep 28, 2026

  228. Qwen Code in VS CodeQwen (Alibaba)Official

    VS Code extension (beta) "Qwen Code Companion" (qwenlm.qwen-code-vscode-ide-companion); JetBrains and Zed through ACP.

    Accessed Sep 28, 2026

  229. Qwen Code settingsQwen (Alibaba)Official

    No default model or reasoning effort set by the agent: the model's own default applies (xhigh for Qwen3.8-Max).

    Accessed Sep 28, 2026

  230. Qwen Code TypeScript SDKQwen (Alibaba)Official

    Qwen Code SDK for TypeScript, with Python and Java SDKs in alpha.

    Accessed Sep 29, 2026

  231. qwen-code-actionQwen (Alibaba)Official

    Official GitHub Action that runs Qwen Code in GitHub workflows.

    Accessed Sep 29, 2026

  232. Qwen3.8-2.4T-A95B model cardQwen (Alibaba)Official

    Vendor coding table for Qwen3.8-Max (Terminal-Bench 2.1 86.6, SWE-bench Pro 67.7, DeepSWE 1.1 56.6) and licence.

    Published Aug 12, 2026 · Accessed Sep 28, 2026

  233. Qwen3.8-Flash-Next model cardQwen (Alibaba)Official

    125B parameters, 6B active, qwen-community-1.0 licence; vendor DeepSWE 58.7, SWE-bench Pro 62.5.

    Published Aug 26, 2026 · Accessed Sep 28, 2026

  234. Deep thinkingQwenCloud (Alibaba)Official

    Qwen3.8-Max reasoning_effort: low, medium, xhigh; xhigh by default.

    Accessed Sep 28, 2026

  235. Qwen3.8-FlashQwenCloud (Alibaba)Official

    $0.15 input / $0.016 cached input / $0.47 output per million tokens.

    Accessed Sep 28, 2026

  236. Qwen3.8-MaxQwenCloud (Alibaba)Official

    $2 input / $0.25 cached input / $6 output per million tokens; 1M-token context.

    Accessed Sep 28, 2026

  237. Token Plan (Team Edition)QwenCloud (Alibaba)Official

    Worked credit example (qwen3.6-plus), from which 1 credit ≈ $0.0025 at API prices is derived for the simulator.

    Accessed Sep 28, 2026

  238. SWE Atlas (codebase Q&A, test writing, refactoring)Scale AIThird party

    Score and confidence interval per model, harness and effort on the three SWE Atlas boards (sweatlas-qna, sweatlas-tw, sweatlas-refactoring); newest entry 2026-09-09.

    Published Sep 9, 2026 · Accessed Sep 29, 2026

  239. SWE-Bench Pro (public) V2Scale AIThird party

    Resolved % of the full and hard subsets per model, harness and effort (update of 2026-09-22: 642 tasks, web tools disabled, re-graded on a clean image; Sonnet 5.5 added 2026-10-12, Gemini 4 Argon and Muse Wonder 1.0 on 2026-10-14, Muse Spark Open on 2026-10-16, MiniMax M3.1 Flash on 2026-11-02, Inkling 2 Small on 2026-11-06). The hard subset is shown.

    Published Nov 6, 2026 · Accessed Nov 11, 2026

  240. Snorkel Agentic Coding 2.0Snorkel AIThird party

    pass@1 and pass@5 with cost per trial and per task for 16 models (benchmark-data/tb_ed4_leaderboard.json loaded by the page; data as_of 2026-10-20, page modified 2026-10-23; no effort stated).

    Published Oct 23, 2026 · Accessed Nov 11, 2026

  241. Snorkel LibraryDesignBench 1.0.0Snorkel AIThird party

    Score (pass rate² × simplicity) with 95% CI, cost per downstream problem and library cost per model × harness; reference arms left out of the layer (the unranked production-library arm scores 46.6).

    Published Oct 1, 2026 · Accessed Nov 11, 2026

  242. Snorkel SWE-bench CLISnorkel AIThird party

    pass@1 and pass@5 with cost per trial and per task for 16 models on multi-file changes (benchmark-data/swe_ed2_leaderboard.json; data as_of 2026-10-20, page modified 2026-10-23; no effort stated).

    Published Oct 23, 2026 · Accessed Nov 11, 2026

  243. Terminal-Bench-Science 0.1 (Snorkel mirror of the official board)Snorkel AI / Laude InstituteThird party

    Resolution rate with CI per model × agent × effort, overall and per science domain, and the total cost of each run (shown rounded); official board at tbench.ai/leaderboard/terminal-bench-science/0.1 (rows loaded client-side).

    Published Sep 24, 2026 · Accessed Nov 11, 2026

  244. Terminal-Bench 4.1Terminal-Bench (tbench.ai)Third party

    Same 66 tasks with hardened verifiers; 4.0 submissions regraded from saved artifacts (162 of 11,550 trials changed, 28 kept at their 4.0 score); every 4.1 score is within its 4.0 95% CI.

    Published Nov 9, 2026 · Accessed Nov 11, 2026

  245. Terminal-Bench 4.1 leaderboardTerminal-Bench (tbench.ai)Third party

    Accuracy with confidence interval, pass@2–5, total cost and tokens per agent × model × effort over 330 trials (Terminal-Bench 4.1, the 2026-11-09 regrade of the 4.0 submissions; the 4.0 page now redirects to the home page).

    Published Nov 9, 2026 · Accessed Nov 11, 2026

  246. Vals AI benchmarksVals AIThird party

    Accuracy, cost per test and latency for Terminal-Bench 4.0 (every model re-run), Vibe Code Bench, Vals Index v2.2, IOI (competitive programming, with standard error), Vibe Code Bench 1-100 and other coding benchmarks (benchmark_view_*.json files linked from each page).

    Published Nov 9, 2026 · Accessed Nov 11, 2026

  247. Vals AI: ProgramBenchVals AIThird party

    Accuracy, cost per test and latency of 66 models (benchmark_view_programbench JSON, metadata.updated 2026-11-09).

    Published Nov 9, 2026 · Accessed Nov 11, 2026

  248. Vals AI: SRE-benchVals AIThird party

    Accuracy, cost per test and latency of 37 models on reverse-engineering real binaries (benchmark_view_sre_bench JSON, metadata.updated 2026-11-05).

    Published Nov 5, 2026 · Accessed Nov 11, 2026

  249. Vals IndexVals AIThird party

    Vals Index v2.2 score, cost per test and latency (Finance Agent v2.1, EMB v1.1, Code Migration v1.1, HLAB-Vals-1.1 and all-pass Tax scoring; scores not comparable with v2.1).

    Accessed Nov 11, 2026

  250. VS Code Marketplace installsVisual Studio MarketplaceThird party

    Total installs on 2026-10-11: GitHub Copilot Chat 78,727,453; GitHub Copilot 74,729,595; Claude Code for VS Code 27,485,072; Codex (openai.chatgpt) 15,360,093; OpenCode (sst-dev.opencode) 1,205,522; Kimi Code 602,491; Google Antigravity (Google.google-antigravity) 557,580; Qwen Code Companion 376,437.

    Accessed Oct 11, 2026

  251. Grok 4.6xAIOfficial

    Coding scores of Grok 4.6 at high effort (DeepSWE v1.1, Terminal-Bench 3.0, CursorBench 3.2).

    Published Aug 12, 2026 · Accessed Sep 28, 2026

  252. Grok 4.7xAIOfficial

    CursorBench 4.0 score, cost per task, output tokens and steps at each effort for Grok 4.7 and for Fable 5.1, Opus 5, Sonnet 5, GPT-5.6 Sol/Terra/Luna and Gemini 3.8 Flash (JS dataset behind the chart).

    Published Sep 21, 2026 · Accessed Sep 28, 2026

  253. Grok 4.7 model cardxAIOfficial

    Terminal-Bench 4.0, FrontierSWE V2 and SWE-Marathon scores of Grok 4.7 and 4.6; CursorBench 4.0 chart (vector geometry) for Grok 4.6.

    Published Sep 21, 2026 · Accessed Sep 28, 2026

  254. Grok BuildxAIOfficial

    Interactive TUI, headless mode and Agent Client Protocol (ACP); command `grok`, installer x.ai/cli/install.sh.

    Accessed Sep 28, 2026

  255. Grok Build headless and ACPxAIOfficial

    IDE integration through ACP (`grok agent stdio`); no dedicated IDE extension.

    Accessed Sep 28, 2026

  256. Grok Build plan modexAIOfficial

    When to plan (architecture ambiguity) and when to skip it (obvious change).

    Accessed Sep 28, 2026

  257. Grok FAQxAIOfficial

    One shared weekly usage pool across Chat, Imagine, Voice and Build since June 2026; shown only as a percentage; extra usage credits.

    Published Aug 27, 2026 · Accessed Sep 28, 2026

  258. Grok plansxAIOfficial

    Monthly and annual prices of SuperGrok Lite, SuperGrok, Plus and Heavy, from the product JSON the page loads (grok.com/rest/products).

    Accessed Sep 28, 2026

  259. ReasoningxAIOfficial

    Effort levels per Grok model (low to xhigh, default high) and guidance per level.

    Accessed Sep 28, 2026

  260. The X Subscription: pricingxAIOfficial

    Free $0, Plus $8, Premium $30, Super $100, Ultra $200 a month (annual $80 / $300 / $1,000 / $2,000), usage 1×, 4×, 15×, 50×; team seats $30 / $100 / $200; existing Cursor, SuperGrok and X Premium plans continue.

    Accessed Oct 17, 2026

  261. xAI API pricingxAIOfficial

    Input, cached input and output prices per million tokens for Grok 4.7, 4.6 and grok-build-0.1.

    Accessed Sep 28, 2026

  262. MiMo CodeXiaomiOfficial

    CLI `mimo` (build, plan and compose modes), MiMo Desktop beta, sign-in with MiMo, Codex or a Claude Code configuration; 13,527 GitHub stars.

    Published Sep 26, 2026 · Accessed Sep 28, 2026

  263. MiMo deep thinkingXiaomiOfficial

    MiMo-V2.6 models: thinking enabled (default) or disabled; no finer effort levels.

    Accessed Sep 28, 2026

  264. MiMo Token PlanXiaomiOfficial

    Lite $6 / Standard $16 / Pro $50 / Max $100 for 4.1B / 11B / 38B / 82B credits per month; no 5-hour or weekly window; credits per token (MiMo-V2.6-Pro 2.5 cache hit / 300 input / 600 output; V2.6-Flash 2 / 100 / 200); off-peak ×0.8.

    Published Sep 21, 2026 · Accessed Sep 28, 2026

  265. MiMo-V2.6-Pro model cardXiaomiOfficial

    1.02T parameters, 42B active, MIT licence, 1M context; vendor Terminal-Bench 4.0 34.9, DeepSWE 1.1 71.9 (not in DeepSWE's official data).

    Published Sep 21, 2026 · Accessed Sep 28, 2026

  266. Claude Code IDE PluginZ.aiOfficial

    Claude Code for VS Code and for JetBrains with the GLM Coding Plan (requires the Claude Code CLI).

    Accessed Sep 28, 2026

  267. GLM Coding PlanZ.aiOfficial

    Prices of Lite, Pro and Max (monthly, quarterly, yearly) and their five-hour and weekly credit allowances.

    Accessed Sep 28, 2026

  268. GLM Coding Plan overviewZ.aiOfficial

    Credit formula and per-model token multipliers, peak and off-peak rates, and the official weekly token estimates per tier.

    Accessed Sep 28, 2026

  269. GLM Coding Plan supported toolsZ.aiOfficial

    Tools allowed with the plan: ZCode, Claude Code, Claude for IDE, Codex, OpenCode, Cline, Kilo Code, Roo Code and others.

    Accessed Sep 28, 2026

  270. GLM-5.3 launch postZ.aiOfficial

    Z.ai Code Bench accuracy and output tokens per task at each effort for GLM-5.3, Fable 5 and Opus 4.8.

    Published Aug 14, 2026 · Accessed Sep 28, 2026

  271. GLM-5.3 model cardZ.aiOfficial

    Coding benchmark table of GLM-5.3 at max effort in Claude Code (Terminal-Bench 2.1 and 3.0, DeepSWE v1.1, NL2Repo, FrontierSWE).

    Accessed Sep 28, 2026

  272. GLM-5.3 model guideZ.aiOfficial

    Effort levels (low, high, max; default max) and the recommendation to use max for coding.

    Accessed Sep 28, 2026

  273. GLM-5.3-FlashZ.aiOfficial

    Coding scores of GLM-5.3-Flash (Terminal-Bench 2.1, DeepSWE v1.1) from printed chart labels.

    Published Aug 26, 2026 · Accessed Sep 28, 2026

  274. Install ZCodeZ.aiOfficial

    ZCode desktop app 3.14.3 for macOS, Windows and Linux; an Agentic Development Environment (ADE).

    Accessed Sep 28, 2026

  275. Z.ai API pricingZ.aiOfficial

    Input, cached input and output prices per million tokens for GLM-5.3 and GLM-5.3-Flash.

    Accessed Sep 28, 2026

  276. ZCode Bot ChannelZ.aiOfficial

    Bot Channel connects ZCode to Feishu and WeChat bots.

    Accessed Sep 29, 2026

  277. ZCode Remote ControlZ.aiOfficial

    Remote Control opens the ZCode desktop window to a phone (QR code or link) to keep instructing the agent.

    Accessed Sep 29, 2026