Calibration
ResearchClawBench → APEX-Agents-AA
APEX-Agents-AA is estimated from ResearchClawBench with a Michaelis–Menten + offset curve fitted on 8 models measured on both: y = 0.0000 + 2.0000·x / (0.89552 + x), R² = 0.32, cross-validated error 10.0 pp. It is used for 3 estimates.
| Estimated model | ResearchClawBench | APEX-Agents-AA | Source |
|---|---|---|---|
| Grok 4.1 | 13.5% | 26.2% | estimated ± 10.0 pp, low confidence |
| MiMo-V2.5 | 16.9% | 31.8% | estimated ± 10.0 pp, low confidence |
| MiMo-V2-Pro | 15.3% | 29.2% | estimated ± 10.0 pp, low confidence |