Banking · Standard · pass^4
τ-bench · 27 models
| Rank | Model | Score |
|---|---|---|
| 1 | Qwen 3.8 Max | 35.0515 |
| 2 | Claude Opus 5 | 31.9588 |
| 2 | Grok 4.5 | 31.9588 |
| 4 | GPT-5.5 | 29.8969 |
| 5 | Claude Fable 5 | 28.866 |
| 6 | GPT-5.6-sol | 27.8351 |
| 7 | Claude Opus 4.7 | 24.7423 |
| 8 | Claude Opus 4.8 | 22.6804 |
| 9 | GPT-5.4 | 21.65 |
| 10 | Muse Spark 1.1 | 20.6186 |
| 11 | GPT-5.2 | 18.56 |
| 12 | Kimi K3 | 17.5258 |
| 13 | GLM-5.2 | 13.4021 |
| 14 | Inkling | 11.3402 |
| 15 | Claude Opus 4.5 | 11.34 |
| 15 | Claude Opus 4.6 | 11.34 |
| 17 | Claude Sonnet 4.5 | 10.31 |
| 18 | Gemini 3.1 Pro Preview | 9.28 |
| 19 | Grok 4.2 | 8.25 |
| 20 | Gemini 3 Flash | 7.22 |
| 21 | Grok 4.1 fast | 5.15 |
| 21 | Qwen3.5-397B-A17B | 5.15 |
| 23 | GPT-5.2 | 4.12 |
| 23 | Gemini 3 Pro | 4.12 |
| 23 | Grok 4 fast | 4.12 |
| 26 | GLM-5 | 3.09 |
| 27 | Gemini 2.5 Pro | 1.03 |