Model Selection Evaluation on HealthBench
90.5Actual (per type)Actual (per type)
Evaluation Results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| Actual (per type)Setting=Pure capability2025.12 | 90.5 | — | — | — | — | — | |
| Actual (per type)Setting=Cost-aware (low)2025.12 | 61.9 | — | — | — | — | — | |
| Actual (per type)Setting=Constrained (single)2025.12 | 53 | — | — | — | — | — | |
| Actual (per type)Setting=Constrained (all)2025.12 | 37.8 | — | — | — | — | — | |
| Actual (per type)Setting=Cost-aware (high)2025.12 | -195.6 | — | — | — | — | — | |
| Actual (mean)Setting=Pure capability2025.12 | — | — | 0.905 | — | — | — | |
| Actual (mean)Setting=Cost-aware (low)2025.12 | — | — | 0.619 | — | — | — | |
| Actual (mean)Setting=Cost-aware (high)2025.12 | — | — | -1.956 | — | — | — | |
| Actual (mean)Setting=Constrained (single)2025.12 | — | — | 0.53 | — | — | — | |
| Actual (mean)Setting=Constrained (all)2025.12 | — | — | 0.378 | — | — | — | |
| BaselinesSetting=Pure capability2025.12 | — | — | — | — | — | 0.53 | |
| BaselinesSetting=Cost-aware (low)2025.12 | — | — | — | — | — | 0.301 | |
| BaselinesSetting=Cost-aware (high)2025.12 | — | — | — | — | — | -1.756 | |
| BaselinesSetting=Constrained (single)2025.12 | — | — | — | — | — | 0.361 | |
| BaselinesSetting=Constrained (all)2025.12 | — | — | — | — | — | 0.306 | |
| MLC (mean)Setting=Pure capability2025.12 | — | — | — | 0.774 | — | — | |
| MLC (mean)Setting=Cost-aware (low)2025.12 | — | — | — | 0.274 | — | — | |
| MLC (mean)Setting=Cost-aware (high)2025.12 | — | — | — | 0.399 | — | — | |
| MLC (mean)Setting=Constrained (single)2025.12 | — | — | — | 0.274 | — | — | |
| MLC (mean)Setting=Constrained (all)2025.12 | — | — | — | 0.274 | — | — | |
| MLC (per type)Setting=Pure capability2025.12 | — | 0.905 | — | — | — | — | |
| MLC (per type)Setting=Cost-aware (low)2025.12 | — | 0.643 | — | — | — | — | |
| MLC (per type)Setting=Cost-aware (high)2025.12 | — | 0.182 | — | — | — | — | |
| MLC (per type)Setting=Constrained (single)2025.12 | — | 0.532 | — | — | — | — | |
| MLC (per type)Setting=Constrained (all)2025.12 | — | 0.378 | — | — | — | — | |
| MLC (robust)Setting=Pure capability2025.12 | — | — | — | — | 0.774 | — | |
| MLC (robust)Setting=Cost-aware (low)2025.12 | — | — | — | — | 0.274 | — | |
| MLC (robust)Setting=Cost-aware (high)2025.12 | — | — | — | — | 0.399 | — | |
| MLC (robust)Setting=Constrained (single)2025.12 | — | — | — | — | 0.274 | — | |
| MLC (robust)Setting=Constrained (all)2025.12 | — | — | — | — | 0.274 | — |