Calibration
Toolathlon-Verified → AA-AnalystAgent
AA-AnalystAgent is estimated from Toolathlon-Verified with a Hill curve fitted on 6 models measured on both: y = 0.0000 + (1.2000 − 0.0000)·x^6.00 / (0.81356^6.00 + x^6.00), R² = 0.65, cross-validated error 6.7 pp. It is used for 3 estimates.
| Estimated model | Toolathlon-Verified | AA-AnalystAgent | Source |
|---|---|---|---|
| Inkling-Small | 54.4% | 9.8% | estimated ± 6.7 pp, low confidence |
| Laguna S 2.1 | 49.7% | 5.9% | estimated ± 6.7 pp, low confidence |
| MiMo-V2.6-Flash | 73.6% | 42.5% | estimated ± 6.7 pp, low confidence |