# HealthBench (raw) → AA-HLE

Calibration · https://benchgap.net/calibration/healthbench/current/to/aa-hle/current · data retrieved 2026-11-21

AA-HLE is estimated from HealthBench (raw) with a Michaelis–Menten curve fitted on 8 models measured on both: y = 1.7872·x / (1.43411 + x), R² = 0.64, cross-validated error 5.1 pp. It is used for 0 estimates.

## Fitted on

| Model | HealthBench (raw) | AA-HLE |
|---|---|---|
| [Claude Opus 5](https://benchgap.net/model/claude-opus-5.md) | 67.1% | 54.9% |
| [Claude Opus 5.5](https://benchgap.net/model/claude-opus-5-5.md) | 68.1% | 61.4% |
| [Claude Sonnet 5.5](https://benchgap.net/model/claude-sonnet-5-5.md) | 69.4% | 55.0% |
| [GPT-6.1 Sol](https://benchgap.net/model/gpt-6-1-sol.md) | 56.7% | 52.9% |
| [GPT-6.1 Terra](https://benchgap.net/model/gpt-6-1-terra.md) | 51.1% | 46.2% |
| [GPT-6 Astra](https://benchgap.net/model/gpt-6-astra.md) | 56.9% | 54.7% |
| [GPT-6 Luna](https://benchgap.net/model/gpt-6-luna.md) | 50.0% | 38.5% |
| [GPT-6 Sol](https://benchgap.net/model/gpt-6-sol.md) | 47.1% | 47.9% |
