CAIA - Pass@1 (With Tools) — leaderboard
Metric: Pass@1 (%). Source: huggingface.co. 17 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | GPT-5 | 70.2 |
| 2 | O3 | 59.6 |
| 3 | Claude Opus 4 | 59.6 |
| 4 | Claude Sonnet 4 | 57.9 |
| 5 | Grok 4 Fast (Reasoning) | 57.9 |
| 6 | Claude Opus 4.1 | 56.3 |
| 7 | GPT-OSS-120B | 56.2 |
| 8 | DeepSeek V3.1 | 55.9 |
| 9 | Grok 4 | 52.2 |
| 10 | GPT-4.1 | 51.7 |
| 11 | GPT-4o | 50 |
| 12 | Gemini 2.5 Pro | 49.4 |
| 13 | Kimi K2 | 46.6 |
| 14 | Gemini 2.5 Flash | 39.3 |
| 15 | Qwen 3 235B A22B | 38.4 |
Interactive version: aibenchmarks.dev/benchmark?slug=caia-pass-1-with-tools · How the rankings work · Data refreshed daily, snapshot 2026-07-12.