Calibration
τ³-bench results → MCP-Tasks
MCP-Tasks is estimated from τ³-bench results with a Hill curve fitted on 5 models measured on both: y = 0.0000 + (1.2000 − 0.0000)·x^6.00 / (0.65083^6.00 + x^6.00), R² = 0.83, cross-validated error 3.6 pp. It is used for 11 estimates.
| Estimated model | τ³-bench results | MCP-Tasks | Source |
|---|---|---|---|
| Atria Dawn Preview | 41.2% | 7.3% | estimated ± 3.6 pp, low confidence |
| Granite 4.2 30B | 62.0% | 51.3% | estimated ± 3.6 pp, low confidence |
| Granite 4.2 3B | 45.8% | 13.0% | estimated ± 3.6 pp, low confidence |
| Granite 4.2 8B | 58.1% | 40.2% | estimated ± 3.6 pp, low confidence |
| LFM2.5-2.6B | 5.7% | 0.0% | estimated ± 3.6 pp, low confidence |
| Mercury 2.5 | 96.0% | 100.0% | estimated ± 3.6 pp, low confidence |
| MiMo-V2.5-Pro | 72.9% | 79.7% | estimated ± 3.6 pp, low confidence |
| Mistral Medium 3.5 128B | 91.4% | 100.0% | estimated ± 3.6 pp, low confidence |
| Nemotron 3.5 Lightning 30B A3B NVFP4 | 9.5% | 0.0% | estimated ± 3.6 pp, low confidence |
| Nemotron 3 Ultra | 70.9% | 75.1% | estimated ± 3.6 pp, low confidence |
| Pokee-Isaac 28B | 66.2% | 63.1% | estimated ± 3.6 pp, medium confidence |