Calibration
MCP Atlas → τ³-bench results
τ³-bench results is estimated from MCP Atlas with a Hill curve fitted on 7 models measured on both: y = 0.4506 + (0.6962 − 0.4506)·x^6.00 / (0.22799^6.00 + x^6.00), R² = 0.63, cross-validated error 1.8 pp. It is used for 31 estimates.
| Estimated model | MCP Atlas | τ³-bench results | Source |
|---|---|---|---|
| Claude Opus 4.7 (Adaptive) | 77.3% | 69.6% | estimated ± 1.8 pp, low confidence |
| Claude Opus 4.8 | 82.2% | 69.6% | estimated ± 1.8 pp, low confidence |
| Claude Opus 5 | 85.8% | 69.6% | estimated ± 1.8 pp, low confidence |
| DeepSeek V4 Flash 0731 | 69.0% | 69.6% | estimated ± 1.8 pp, medium confidence |
| DeepSeek V4 Pro 0813 | 73.6% | 69.6% | estimated ± 1.8 pp, low confidence |
| Gemini 3.5 Flash | 83.6% | 69.6% | estimated ± 1.8 pp, low confidence |
| GLM-5.2 | 76.8% | 69.6% | estimated ± 1.8 pp, low confidence |
| GPT-5.4 | 70.6% | 69.6% | estimated ± 1.8 pp, medium confidence |
| GPT-5.4 mini | 57.7% | 69.5% | estimated ± 1.8 pp, medium confidence |
| GPT-5.4 nano | 56.1% | 69.5% | estimated ± 1.8 pp, medium confidence |
| GPT-5.5 | 75.3% | 69.6% | estimated ± 1.8 pp, low confidence |
| Hy4 preview | 83.7% | 69.6% | estimated ± 1.8 pp, low confidence |
| Inkling | 74.1% | 69.6% | estimated ± 1.8 pp, low confidence |
| Inkling-Small | 79.6% | 69.6% | estimated ± 1.8 pp, low confidence |
| Kimi K2.6 | 55.9% | 69.5% | estimated ± 1.8 pp, medium confidence |
| Kimi K2.7 Code | 76.0% | 69.6% | estimated ± 1.8 pp, low confidence |
| Kimi K3 | 84.2% | 69.6% | estimated ± 1.8 pp, low confidence |
| Ling 3.0 Flash | 65.5% | 69.6% | estimated ± 1.8 pp, medium confidence |
| LLaDA2.2-flash | 46.2% | 69.3% | estimated ± 1.8 pp, medium confidence |
| LongCat-Flash-Lite-Sparse | 45.6% | 69.2% | estimated ± 1.8 pp, medium confidence |
| MiniMax M3 | 74.2% | 69.6% | estimated ± 1.8 pp, low confidence |
| Muse Glimmer 30B | 75.5% | 69.6% | estimated ± 1.8 pp, low confidence |
| Muse Spark 1.1 | 88.1% | 69.6% | estimated ± 1.8 pp, low confidence |
| Ornith-1.5-35B-A3B | 70.2% | 69.6% | estimated ± 1.8 pp, medium confidence |
| Ornith-1.5-397B | 80.0% | 69.6% | estimated ± 1.8 pp, low confidence |
| Ornith-1.5-9B | 54.2% | 69.5% | estimated ± 1.8 pp, medium confidence |
| Qwen3.7 Max | 76.4% | 69.6% | estimated ± 1.8 pp, low confidence |
| Beam | 78.7% | 69.6% | estimated ± 1.8 pp, low confidence |
| Solar Open 2 | 58.2% | 69.5% | estimated ± 1.8 pp, medium confidence |
| Solar Pro 4 | 61.4% | 69.6% | estimated ± 1.8 pp, medium confidence |
| Step 5 Preview | 85.6% | 69.6% | estimated ± 1.8 pp, low confidence |