ModelCompare
Compare

AI coding agent usage

What was the most used AI coding agent on ?

Claude Code from Anthropic, with a usage score of 70 out of 100 across 5 of 5 sources. GitHub Copilot was next at 57.

Usage score over time, week by week

Usage ranking

AI coding agents by usage on September 1, 2026

This ranks how much each agent is used and followed, not how good it is: a popular agent can be a worse fit for your work than a small one. Click an agent to see where its numbers come from. Each source's column shows the agent's scaled score (the leader is 100) over the raw number. 5 sources had data on this day. The score averages an agent's scaled values over the sources that can measure it (at least three).

  1. 1
    Claude Code
    Anthropic · closed source
    Usage score 70
    • npm 83 · 19M
    • OpenRouter 100 · 14T
    • GitHub 37 · 144k
    • VS Code 29 · 22M
    • JetBrains 100 · 39%
    • npm: 8318,969,469 downloads of @anthropic-ai/claude-code in the 7 days to Sep 1, 2026. The leader here is Codex CLI with 23M, so Claude Code scores 19M ÷ 23M × 100 = 83. Source
    • OpenRouter: 10014T tokens sent through OpenRouter by Claude Code in the week to Sep 2, 2026. That's the most of any agent in this source, so it scores 100. Source
    • GitHub: 37143,664 stars on github.com/anthropics/claude-code on Sep 1, 2026. The leader here is OpenClaw with 388k, so Claude Code scores 144k ÷ 388k × 100 = 37. Source
    • VS Code: 2922,430,605 installs of the anthropic.claude-code extension on Aug 5, 2026. The leader here is GitHub Copilot with 77M, so Claude Code scores 22M ÷ 77M × 100 = 29. Source
    • JetBrains Developer Ecosystem 2026: 10039% of 15,000 respondents said they use it. Question: AI coding tools used at work, May to July 2026. That's the most of any agent in this source, so it scores 100. Source

    Usage score: (83 + 100 + 37 + 29 + 100) ÷ 5 = 70. Claude Code website

  2. 2
    GitHub Copilot
    GitHub · closed source
    Usage score 57
    • npm 18 · 4.2M
    • VS Code 100 · 77M
    • JetBrains 54 · 21%
    • npm: 184,216,105 downloads of @github/copilot in the 7 days to Sep 1, 2026. The leader here is Codex CLI with 23M, so GitHub Copilot scores 4.2M ÷ 23M × 100 = 18. Source
    • VS Code: 10077,249,711 installs of the GitHub.copilot-chat extension on Aug 4, 2026. That's the most of any agent in this source, so it scores 100. Source
    • JetBrains Developer Ecosystem 2026: 5421% of 15,000 respondents said they use it. Question: AI coding tools used at work, May to July 2026. The leader here is Claude Code with 39%, so GitHub Copilot scores 21% ÷ 39% × 100 = 54. Source

    Not counted: OpenRouter, because it doesn't show up in OpenRouter's app rankings; GitHub, because it has no public GitHub repo.

    Usage score: (18 + 100 + 54) ÷ 3 = 57. GitHub Copilot website

  3. 3
    OpenClaw
    OpenClaw · open source
    Usage score 49
    • npm 15 · 3.3M
    • OpenRouter 33 · 4.6T
    • GitHub 100 · 388k
    • npm: 153,343,010 downloads of openclaw in the 7 days to Sep 1, 2026. The leader here is Codex CLI with 23M, so OpenClaw scores 3.3M ÷ 23M × 100 = 15. Source
    • OpenRouter: 334.6T tokens sent through OpenRouter by OpenClaw in the week to Sep 2, 2026. The leader here is Claude Code with 14T, so OpenClaw scores 4.6T ÷ 14T × 100 = 33. Source
    • GitHub: 100388,310 stars on github.com/openclaw/openclaw on Sep 1, 2026. That's the most of any agent in this source, so it scores 100. Source

    Not counted: VS Code, because it has no VS Code extension; JetBrains Developer Ecosystem 2026, because the survey didn't list it.

    Usage score: (15 + 33 + 100) ÷ 3 = 49. OpenClaw website

  4. 4
    Codex CLI
    OpenAI · open source
    Usage score 41
    • npm 100 · 23M
    • OpenRouter 18 · 2.5T
    • GitHub 31 · 121k
    • VS Code 18 · 14M
    • JetBrains 41 · 16%
    • npm: 10022,794,236 downloads of @openai/codex in the 7 days to Sep 1, 2026. That's the most of any agent in this source, so it scores 100. Source
    • OpenRouter: 182.5T tokens sent through OpenRouter by Codex CLI in the week to Sep 2, 2026. The leader here is Claude Code with 14T, so Codex CLI scores 2.5T ÷ 14T × 100 = 18. Source
    • GitHub: 31120,662 stars on github.com/openai/codex on Sep 1, 2026. The leader here is OpenClaw with 388k, so Codex CLI scores 121k ÷ 388k × 100 = 31. Source
    • VS Code: 1813,722,411 installs of the openai.chatgpt extension on Sep 1, 2026. The leader here is GitHub Copilot with 77M, so Codex CLI scores 14M ÷ 77M × 100 = 18. Source
    • JetBrains Developer Ecosystem 2026: 4116% of 15,000 respondents said they use it. Question: AI coding tools used at work, May to July 2026. The leader here is Claude Code with 39%, so Codex CLI scores 16% ÷ 39% × 100 = 41. Source

    Usage score: (100 + 18 + 31 + 18 + 41) ÷ 5 = 41. Codex CLI website

  5. 5
    OpenCode
    Anomaly · open source
    Usage score 21
    • npm 10 · 2.2M
    • GitHub 52 · 203k
    • VS Code 1 · 1.1M
    • npm: 102,165,567 downloads of opencode-ai in the 7 days to Sep 1, 2026. The leader here is Codex CLI with 23M, so OpenCode scores 2.2M ÷ 23M × 100 = 10. Source
    • GitHub: 52202,872 stars on github.com/anomalyco/opencode on Sep 1, 2026. The leader here is OpenClaw with 388k, so OpenCode scores 203k ÷ 388k × 100 = 52. Source
    • VS Code: 11,059,402 installs of the sst-dev.opencode extension on Sep 1, 2026. The leader here is GitHub Copilot with 77M, so OpenCode scores 1.1M ÷ 77M × 100 = 1. Source

    Not counted: OpenRouter, because it doesn't show up in OpenRouter's app rankings; JetBrains Developer Ecosystem 2026, because the survey didn't list it.

    Usage score: (10 + 52 + 1) ÷ 3 = 21. OpenCode website

  6. 6
    Cline
    Cline · open source
    Usage score 21
    • npm 0 · 86k
    • OpenRouter 59 · 8.3T
    • GitHub 17 · 65k
    • VS Code 6 · 4.9M
    • npm: 085,995 downloads of cline in the 7 days to Sep 1, 2026. The leader here is Codex CLI with 23M, so Cline scores 86k ÷ 23M × 100 = 0. Source
    • OpenRouter: 598.3T tokens sent through OpenRouter by Cline in the week to Sep 2, 2026. The leader here is Claude Code with 14T, so Cline scores 8.3T ÷ 14T × 100 = 59. Source
    • GitHub: 1765,354 stars on github.com/cline/cline on Aug 1, 2026. The leader here is OpenClaw with 388k, so Cline scores 65k ÷ 388k × 100 = 17. Source
    • VS Code: 64,863,603 installs of the saoudrizwan.claude-dev extension on Aug 3, 2026. The leader here is GitHub Copilot with 77M, so Cline scores 4.9M ÷ 77M × 100 = 6. Source

    Not counted: JetBrains Developer Ecosystem 2026, because the survey didn't list it.

    Usage score: (0 + 59 + 17 + 6) ÷ 4 = 21. Cline website

  7. 7
    Hermes Agent
    Nous Research · open source
    Usage score 21
    • OpenRouter 0 · 6.5B
    • GitHub 62 · 239k
    • OpenRouter: 06.5B tokens sent through OpenRouter by Hermes Agent in the week to Sep 2, 2026. The leader here is Claude Code with 14T, so Hermes Agent scores 6.5B ÷ 14T × 100 = 0. Source
    • GitHub: 62239,371 stars on github.com/NousResearch/hermes-agent on Sep 1, 2026. The leader here is OpenClaw with 388k, so Hermes Agent scores 239k ÷ 388k × 100 = 62. Source

    Not counted: npm, because it has no command-line package on npm; VS Code, because it has no VS Code extension; JetBrains Developer Ecosystem 2026, because the survey didn't list it.

    Usage score: (0 + 62) ÷ 3 = 21, divided by 3 rather than 2 because every agent's average counts at least 3 sources. Hermes Agent website

  8. 8
    Kilo Code
    Kilo · open source
    Usage score 19
    • npm 0 · 25k
    • OpenRouter 66 · 9.3T
    • GitHub 7 · 27k
    • VS Code 2 · 1.5M
    • npm: 025,485 downloads of @kilocode/cli in the 7 days to Sep 1, 2026. The leader here is Codex CLI with 23M, so Kilo Code scores 25k ÷ 23M × 100 = 0. Source
    • OpenRouter: 669.3T tokens sent through OpenRouter by Kilo Code in the week to Sep 2, 2026. The leader here is Claude Code with 14T, so Kilo Code scores 9.3T ÷ 14T × 100 = 66. Source
    • GitHub: 726,698 stars on github.com/Kilo-Org/kilocode on Aug 4, 2026. The leader here is OpenClaw with 388k, so Kilo Code scores 27k ÷ 388k × 100 = 7. Source
    • VS Code: 21,479,528 installs of the kilocode.Kilo-Code extension on Aug 30, 2026. The leader here is GitHub Copilot with 77M, so Kilo Code scores 1.5M ÷ 77M × 100 = 2. Source

    Not counted: JetBrains Developer Ecosystem 2026, because the survey didn't list it.

    Usage score: (0 + 66 + 7 + 2) ÷ 4 = 19. Kilo Code website

  9. 9
    pi
    Mario Zechner · open source
    Usage score 13
    • npm 0 · 7.6k
    • OpenRouter 39 · 5.5T
    • npm: 07,588 downloads of @mariozechner/pi-coding-agent in the 7 days to Sep 1, 2026. The leader here is Codex CLI with 23M, so pi scores 7.6k ÷ 23M × 100 = 0. Source
    • OpenRouter: 395.5T tokens sent through OpenRouter by pi in the week to Sep 2, 2026. The leader here is Claude Code with 14T, so pi scores 5.5T ÷ 14T × 100 = 39. Source

    Not counted: GitHub, because there's no snapshot of its repo from around that day; VS Code, because it has no VS Code extension; JetBrains Developer Ecosystem 2026, because the survey didn't list it.

    Usage score: (0 + 39) ÷ 3 = 13, divided by 3 rather than 2 because every agent's average counts at least 3 sources. pi website

  10. 10
    Gemini CLI
    Google · open source
    Usage score 10
    • npm 2 · 391k
    • GitHub 27 · 106k
    • npm: 2390,652 downloads of @google/gemini-cli in the 7 days to Sep 1, 2026. The leader here is Codex CLI with 23M, so Gemini CLI scores 391k ÷ 23M × 100 = 2. Source
    • GitHub: 27106,320 stars on github.com/google-gemini/gemini-cli on Aug 3, 2026. The leader here is OpenClaw with 388k, so Gemini CLI scores 106k ÷ 388k × 100 = 27. Source

    Not counted: OpenRouter, because it doesn't show up in OpenRouter's app rankings; VS Code, because there's no snapshot of its extension page from around that day; JetBrains Developer Ecosystem 2026, because the survey didn't list it.

    Usage score: (2 + 27) ÷ 3 = 10, divided by 3 rather than 2 because every agent's average counts at least 3 sources. Gemini CLI website

  11. 11
    OpenHands
    All Hands AI · open source
    Usage score 9
    • OpenRouter 5 · 716B
    • GitHub 21 · 83k
    • OpenRouter: 5716B tokens sent through OpenRouter by OpenHands in the week to Sep 2, 2026. The leader here is Claude Code with 14T, so OpenHands scores 716B ÷ 14T × 100 = 5. Source
    • GitHub: 2182,903 stars on github.com/OpenHands/OpenHands on Aug 2, 2026. The leader here is OpenClaw with 388k, so OpenHands scores 83k ÷ 388k × 100 = 21. Source

    Not counted: npm, because it has no command-line package on npm; VS Code, because it has no VS Code extension; JetBrains Developer Ecosystem 2026, because the survey didn't list it.

    Usage score: (5 + 21) ÷ 3 = 9, divided by 3 rather than 2 because every agent's average counts at least 3 sources. OpenHands website

  12. 12
    Aider
    Aider · open source
    Usage score 4
    • OpenRouter 0 · 0
    • GitHub 13 · 49k
    • OpenRouter: 00 tokens sent through OpenRouter by Aider in the week to Sep 2, 2026. The leader here is Claude Code with 14T, so Aider scores 0 ÷ 14T × 100 = 0. Source
    • GitHub: 1348,648 stars on github.com/Aider-AI/aider on Sep 1, 2026. The leader here is OpenClaw with 388k, so Aider scores 49k ÷ 388k × 100 = 13. Source

    Not counted: npm, because it has no command-line package on npm; VS Code, because it has no VS Code extension; JetBrains Developer Ecosystem 2026, because the survey didn't list it.

    Usage score: (0 + 13) ÷ 3 = 4, divided by 3 rather than 2 because every agent's average counts at least 3 sources. Aider website

  13. 13
    Roo Code
    Roo Code · open source
    Usage score 3
    • OpenRouter 2 · 213B
    • GitHub 6 · 24k
    • VS Code 3 · 1.9M
    • OpenRouter: 2213B tokens sent through OpenRouter by Roo Code in the week to Sep 2, 2026. The leader here is Claude Code with 14T, so Roo Code scores 213B ÷ 14T × 100 = 2. Source
    • GitHub: 624,337 stars on github.com/RooCodeInc/Roo-Code on Aug 16, 2026. The leader here is OpenClaw with 388k, so Roo Code scores 24k ÷ 388k × 100 = 6. Source
    • VS Code: 31,940,604 installs of the RooVeterinaryInc.roo-cline extension on Aug 20, 2026. The leader here is GitHub Copilot with 77M, so Roo Code scores 1.9M ÷ 77M × 100 = 3. Source

    Not counted: npm, because it has no command-line package on npm; JetBrains Developer Ecosystem 2026, because the survey didn't list it.

    Usage score: (2 + 6 + 3) ÷ 3 = 3. Roo Code website

  14. 14
    Qwen Code
    Alibaba · open source
    Usage score 3
    • npm 0 · 73k
    • OpenRouter 1 · 159B
    • GitHub 7 · 27k
    • npm: 072,999 downloads of @qwen-code/qwen-code in the 7 days to Sep 1, 2026. The leader here is Codex CLI with 23M, so Qwen Code scores 73k ÷ 23M × 100 = 0. Source
    • OpenRouter: 1159B tokens sent through OpenRouter by Qwen Code in the week to Sep 2, 2026. The leader here is Claude Code with 14T, so Qwen Code scores 159B ÷ 14T × 100 = 1. Source
    • GitHub: 727,216 stars on github.com/QwenLM/qwen-code on Aug 20, 2026. The leader here is OpenClaw with 388k, so Qwen Code scores 27k ÷ 388k × 100 = 7. Source

    Not counted: VS Code, because it has no VS Code extension; JetBrains Developer Ecosystem 2026, because the survey didn't list it.

    Usage score: (0 + 1 + 7) ÷ 3 = 3. Qwen Code website

  15. 15
    Crush
    Charm · open source
    Usage score 2
    • npm 0 · 8k
    • GitHub 7 · 27k
    • npm: 08,032 downloads of @charmland/crush in the 7 days to Sep 1, 2026. The leader here is Codex CLI with 23M, so Crush scores 8k ÷ 23M × 100 = 0. Source
    • GitHub: 727,248 stars on github.com/charmbracelet/crush on Aug 10, 2026. The leader here is OpenClaw with 388k, so Crush scores 27k ÷ 388k × 100 = 7. Source

    Not counted: OpenRouter, because it doesn't show up in OpenRouter's app rankings; VS Code, because it has no VS Code extension; JetBrains Developer Ecosystem 2026, because the survey didn't list it.

    Usage score: (0 + 7) ÷ 3 = 2, divided by 3 rather than 2 because every agent's average counts at least 3 sources. Crush website

  16. 16
    Letta Code
    Letta · open source
    Usage score 0
    • npm 0 · 67k
    • OpenRouter 0 · 56B
    • GitHub 1 · 3.1k
    • npm: 066,640 downloads of @letta-ai/letta-code in the 7 days to Sep 1, 2026. The leader here is Codex CLI with 23M, so Letta Code scores 67k ÷ 23M × 100 = 0. Source
    • OpenRouter: 056B tokens sent through OpenRouter by Letta Code in the week to Sep 2, 2026. The leader here is Claude Code with 14T, so Letta Code scores 56B ÷ 14T × 100 = 0. Source
    • GitHub: 13,116 stars on github.com/letta-ai/letta-code on Aug 25, 2026. The leader here is OpenClaw with 388k, so Letta Code scores 3.1k ÷ 388k × 100 = 1. Source

    Not counted: VS Code, because it has no VS Code extension; JetBrains Developer Ecosystem 2026, because the survey didn't list it.

    Usage score: (0 + 0 + 1) ÷ 3 = 0. Letta Code website

Not ranked that day, only one source has them: Cursor (JetBrains, 12%), Auggie (npm, 30k), Amp (npm, 19k), Droid (npm, 9.5k).

Method

How the usage score is made

No single public number measures how much each AI coding agent is used, so we blend the ones that exist. In each source the leader scores 100 and the rest are scaled to it. An agent's score is the average over the sources that can measure it. Sources that can't see an agent at all (Cursor has no npm package) are left out of its average, but the average always counts at least three sources, so an agent seen by only one or two can't outrank one that leads everywhere. An agent needs two sources to be ranked.

  • npmCLI downloads, 7 days · data from Sep 1, 2026

    Downloads of each agent's command-line package in the 7 days to the date, with one-off bursts (mirrors, bots) capped. Downloads count installs and CI runs, not people, and agents with their own installer are undercounted. Source

  • OpenRouterTokens via OpenRouter, 7 days · data from Sep 2, 2026

    Tokens each app sent through OpenRouter in a week. It only sees traffic routed through OpenRouter, so agents that mostly call their own lab's API (Claude Code, Codex) look smaller than they are, and open-source agents that default to OpenRouter look bigger. Source

  • GitHubStars · data from Sep 1, 2026

    Stars on the agent's GitHub repo, from monthly snapshots. Stars measure interest, and build up over time. Source

  • VS CodeExtension installs, all time · data from Sep 1, 2026

    All-time installs of the agent's VS Code extension, from monthly snapshots. Cursor and Windsurf install from Open VSX instead, which isn't counted. Source

  • JetBrains Developer Ecosystem 2026% of 15,000 respondents · data from Jul 31, 2026

    Share of developers in the survey who said they use each tool. Surveys only list the biggest tools. Source

Browse

The most used AI coding agent, month by month

2026
2025

Pick the model behind your agent

Most of these agents let you choose the model. Send one prompt to up to six models and see which answer you'd actually use, with what each one cost.

Compare models

Data through October 21, 2026 for 20 agents. npm downloads from the npm registry; OpenRouter app rankings from OpenRouter (CC BY 4.0); GitHub stars and VS Code installs from monthly snapshots in the Internet Archive and the live APIs; survey figures as published by Stack Overflow and JetBrains.