claude code vs cursor vs codex — measured, not surveyed
| # | Agent | Devs | Tokens | Share | Top model | Est cost |
|---|---|---|---|---|---|---|
| 01 | Codex CLIcodex | 93 | 465.62B | 62% | gpt-5.6-sol | $219,212 |
| 02 | Claude Codeclaude | 96 | 245.36B | 33% | claude-opus-5 | $234,463 |
| 03 | OpenCodeopencode | 40 | 16.38B | 2.2% | deepseek-v4-flash-free | $2,985 |
| 04 | Cursorcursor | 32 | 14.64B | 2.0% | default | $12,419 |
| 05 | Hermeshermes | 18 | 2.62B | 0.3% | gpt-5.6-sol | $1,026 |
| 06 | Kimi CLIkimi | 4 | 2.36B | 0.3% | k3 | $830 |
| 07 | Pipi | 5 | 354.7M | <0.1% | model-237bf97a0500 | $16.6 |
| 08 | Factory Droiddroid | 2 | 88.2M | <0.1% | glm-5-2-fast | $0.4 |
| 09 | Kilo Codekilo | 3 | 7.0M | <0.1% | free | $0 |
| 10 | Gemini CLIgemini | 2 | 1.7M | <0.1% | gemini-2.5-pro | $0 |
| 11 | Clinecline | 2 | 1.2M | <0.1% | deepseek-v4-flash | $0.26 |
npx whoburnedmore CLI reads the local usage logs coding agents already keep and submits daily aggregate token totals per tool and model — never prompts, code or file names. Panel = 144 listed developers who opted in publicly, including subscription usage (Claude Code, Cursor, Codex) that API-gateway rankings never see. Costs estimated from public per-token pricing. what we collect · leaderboard · state of AI coding