CAIA - Pass@1 (Without Tools) — leaderboard
Metric: Pass@1 (%). Source: huggingface.co. 17 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | GPT-5 | 28.1 |
| 2 | O3 | 22.5 |
| 3 | DeepSeek R1 | 21.9 |
| 4 | Grok 4 Fast (Reasoning) | 21.3 |
| 5 | GPT-4.1 | 20.8 |
| 6 | Gemini 2.5 Pro | 20.2 |
| 7 | Gemini 2.5 Flash | 20.2 |
| 8 | Grok 4 | 20.2 |
| 9 | GPT-4o | 19.1 |
| 10 | GPT-OSS-120B | 18.5 |
| 11 | Kimi K2 | 17.4 |
| 12 | Llama 4 Maverick | 16.9 |
| 13 | DeepSeek V3.1 | 15.7 |
| 14 | Claude Opus 4.1 | 15.2 |
| 15 | Qwen 3 235B A22B | 14.6 |
Interactive version: aibenchmarks.dev/benchmark?slug=caia-pass-1-without-tools · How the rankings work · Data refreshed daily, snapshot 2026-07-12.