Banking · Standard · pass^3
τ-bench · 27 models
| Rank | Model | Score |
|---|---|---|
| 1 | Qwen 3.8 Max | 39.6907 |
| 2 | Grok 4.5 | 35.3093 |
| 3 | Claude Opus 5 | 34.5361 |
| 4 | GPT-5.5 | 33.2474 |
| 5 | GPT-5.6-sol | 31.9588 |
| 6 | Claude Fable 5 | 30.1546 |
| 7 | Claude Opus 4.7 | 27.0619 |
| 8 | Claude Opus 4.8 | 26.2887 |
| 9 | GPT-5.4 | 24.48 |
| 10 | Muse Spark 1.1 | 23.9691 |
| 11 | GPT-5.2 | 20.62 |
| 12 | Kimi K3 | 20.1031 |
| 13 | GLM-5.2 | 18.299 |
| 14 | Claude Opus 4.6 | 14.18 |
| 15 | Claude Opus 4.5 | 13.66 |
| 16 | Inkling | 13.4021 |
| 17 | Claude Sonnet 4.5 | 12.63 |
| 18 | Gemini 3.1 Pro Preview | 11.08 |
| 19 | Gemini 3 Flash | 10.31 |
| 20 | Grok 4.2 | 10.05 |
| 21 | GLM-5 | 5.67 |
| 21 | Gemini 3 Pro | 5.67 |
| 21 | Grok 4 fast | 5.67 |
| 21 | Grok 4.1 fast | 5.67 |
| 25 | Qwen3.5-397B-A17B | 5.41 |
| 26 | GPT-5.2 | 5.15 |
| 27 | Gemini 2.5 Pro | 2.58 |