Banking · Standard · pass^2
τ-bench · 27 models
| Rank | Model | Score |
|---|---|---|
| 1 | Qwen 3.8 Max | 45.189 |
| 2 | Grok 4.5 | 39.5189 |
| 3 | Claude Opus 5 | 39.1753 |
| 4 | GPT-5.5 | 37.6289 |
| 5 | GPT-5.6-sol | 37.2852 |
| 6 | Claude Fable 5 | 32.8179 |
| 7 | Claude Opus 4.8 | 31.2715 |
| 8 | Claude Opus 4.7 | 31.0997 |
| 9 | Muse Spark 1.1 | 29.5533 |
| 10 | GPT-5.4 | 29.55 |
| 11 | GLM-5.2 | 25.2577 |
| 11 | Kimi K3 | 25.2577 |
| 13 | GPT-5.2 | 23.88 |
| 14 | Claude Opus 4.6 | 18.73 |
| 15 | Claude Opus 4.5 | 17.18 |
| 16 | Inkling | 16.6667 |
| 17 | Claude Sonnet 4.5 | 15.81 |
| 18 | Gemini 3 Flash | 15.46 |
| 19 | Gemini 3.1 Pro Preview | 15.12 |
| 20 | Grok 4.2 | 12.37 |
| 21 | Gemini 3 Pro | 9.28 |
| 22 | Grok 4 fast | 8.42 |
| 23 | Grok 4.1 fast | 7.73 |
| 24 | GLM-5 | 7.22 |
| 24 | GPT-5.2 | 7.22 |
| 26 | Qwen3.5-397B-A17B | 6.36 |
| 27 | Gemini 2.5 Pro | 6.01 |