Community usage

Which AI coding agents and models developers actually use

Every agent launched from AgentsRoom by users who share their usage, counted by CLI and by model. No survey, no vendor benchmark: what people start when they have work to do.

Updated Oct 15, 2026, 1:32 PM

Agent launches

100,917

newvs previous period

Users

784

CLIs used

17

Models used

60

Most launched coding CLIs

Share of all agent launches over the period, with the users who launched each one at least once.

  1. 1
    Claude59.1%
    59665 launches706 usersnew
  2. 2
    Codex28.1%
    28384 launches258 usersnew
  3. 3
    Antigravity4.2%
    4217 launches126 usersnew
  4. 4
    OpenCode3.7%
    3695 launches112 usersnew
  5. 5
    Mistral Vibe1.4%
    1445 launches5 usersnew
  6. 6
    GitHub Copilot0.8%
    826 launches19 usersnew
  7. 7
    Grok0.6%
    593 launches26 usersnew
  8. 8
    oh-my-pi0.5%
    555 launches27 usersnew
  9. 9
    Cursor0.5%
    470 launches27 usersnew
  10. 10
    Ollama0.3%
    336 launches34 usersnew
  11. 11
    Freebuff0.2%
    231 launches40 usersnew
  12. 12
    Kimi0.2%
    224 launches14 usersnew
  13. 13
    Devin0.2%
    162 launches11 usersnew
  14. 14
    Aider0.1%
    106 launches9 usersnew
  15. 15
    Gemini0%
    4 launches1 usersnew
  16. 16
    ANTIGRAVITY0%
    3 launches0 usersnew
  17. 17
    Amp0%
    1 launches0 usersnew

Most launched models

The model pinned on the agent at launch. Launches that leave the model to the CLI's default are counted in the CLI totals only.

  1. 1
    OpusClaude21.9%
    22150 launches654 users37.1% of Claude launchesnew
  2. 2
    SonnetClaude12.2%
    12283 launches247 users20.6% of Claude launchesnew
  3. 3
    Opus 5.5Claude9.9%
    10008 launches143 users16.8% of Claude launchesnew
  4. 4
    GPT-6.1 SolCodex8.8%
    8919 launches152 users31.4% of Codex launchesnew
  5. 5
    GPT-6 LunaCodex5.2%
    5274 launches57 users18.6% of Codex launchesnew
  6. 6
    GPT-5.6 SolCodex4.1%
    4174 launches127 users14.7% of Codex launchesnew
  7. 7
    GPT-6 AstraCodex3.4%
    3437 launches85 users12.1% of Codex launchesnew
  8. 8
    GPT-6 SolCodex3.1%
    3131 launches59 users11% of Codex launchesnew
  9. 9
    HaikuClaude2.7%
    2765 launches89 users4.6% of Claude launchesnew
  10. 10
    FableClaude2.3%
    2342 launches80 users3.9% of Claude launchesnew
  11. 11
    defaultClaude2.1%
    2105 launches177 users3.5% of Claude launchesnew
  12. 12
    Sonnet 5.5Claude1.9%
    1938 launches66 users3.2% of Claude launchesnew
  13. 13
    Medium 3.5Mistral Vibe1.4%
    1444 launches4 users99.9% of Mistral Vibe launchesnew
  14. 14
    GPT-5.6 TerraCodex1.4%
    1364 launches34 users4.8% of Codex launchesnew
  15. 15
    Gemini 3.1 Pro (High)Antigravity1.2%
    1212 launches92 users28.7% of Antigravity launchesnew
  16. 16
    Gemini 3.8 Flash (High)Antigravity1.1%
    1088 launches48 users25.8% of Antigravity launchesnew
  17. 17
    Opus 5Claude1%
    996 launches51 users1.7% of Claude launchesnew
  18. 18
    gemini-3.8-flash-lowAntigravity0.8%
    760 launches3 users18% of Antigravity launchesnew
  19. 19
    GPT-5.6 LunaCodex0.7%
    727 launches24 users2.6% of Codex launchesnew
  20. 20
    claude-opus-5-5Claude0.5%
    462 launches6 users0.8% of Claude launchesnew

And 40 more models below the cut.

How agents are started

  • New agent25%
  • Trigger21%
  • Team run16%
  • Backlog ticket12%
  • Resumed12%
  • Quick agent9%

Reasoning effort

  • High39%
  • Medium30%
  • Low14%
  • Max8%
  • Xhigh6%
  • Ultra2%

Agent roles

  • Custom25%
  • Fullstack22%
  • Pm13%
  • Backend7%
  • Qa7%
  • Architect6%

Where the numbers come from

AgentsRoom is a desktop command center for AI coding agents. Every time a user starts an agent, the desktop records the CLI, the model, the reasoning effort, the agent role and how it was started, and nothing else: no prompt, no project, no file, no name. Only users who keep "Share usage data" on contribute, anonymous launches count as launches but never as users, and the page refreshes every ten minutes.

The same numbers rank the Providers and Models boards of the in-app leaderboard, put a flame on the three most launched models of each provider and drive the Popular sort of every model list. How the model picker uses them

Looking for the model of another CLI that plays the same role as yours? AI model equivalents, tier by tier

Download AgentsRoom