Banking · Standard · pass^1
τ-bench · 27 models
| Rank | Model | Score |
|---|---|---|
| 1 | Qwen 3.8 Max | 55.1546 |
| 2 | Claude Opus 5 | 48.7113 |
| 3 | Grok 4.5 | 47.9381 |
| 4 | GPT-5.6-sol | 46.9072 |
| 5 | GPT-5.5 | 44.5876 |
| 6 | Muse Spark 1.1 | 40.4639 |
| 7 | Claude Opus 4.7 | 40.2062 |
| 8 | Claude Fable 5 | 39.6907 |
| 8 | Claude Opus 4.8 | 39.6907 |
| 10 | GPT-5.4 | 39.43 |
| 11 | GLM-5.2 | 37.1134 |
| 11 | Kimi K3 | 37.1134 |
| 13 | GPT-5.2 | 32.22 |
| 14 | Claude Opus 4.6 | 27.32 |
| 14 | Gemini 3 Flash | 27.32 |
| 16 | Gemini 3.1 Pro Preview | 26.03 |
| 17 | Claude Sonnet 4.5 | 25.26 |
| 18 | Inkling | 25 |
| 19 | Claude Opus 4.5 | 24.74 |
| 20 | Gemini 3 Pro | 18.04 |
| 20 | Grok 4.2 | 18.04 |
| 22 | Grok 4 fast | 15.72 |
| 23 | Gemini 2.5 Pro | 13.66 |
| 24 | Grok 4.1 fast | 13.14 |
| 25 | GPT-5.2 | 12.63 |
| 26 | GLM-5 | 9.79 |
| 26 | Qwen3.5-397B-A17B | 9.79 |