Calibration
FrontierCode 1.1 Main → AA Coding Index
AA Coding Index is estimated from FrontierCode 1.1 Main with a Michaelis–Menten curve fitted on 9 models measured on both: y = 1.1662·x / (0.25924 + x), R² = 0.86, cross-validated error 4.4 pp. It is used for 1 estimate.
| Estimated model | FrontierCode 1.1 Main | AA Coding Index | Source |
|---|---|---|---|
| Claude Opus 4.6 | 26.9% | 59.4% | estimated ± 4.4 pp, medium confidence |
Fitted on
| Model | FrontierCode 1.1 Main | AA Coding Index |
|---|---|---|
| Claude Fable 5 | 53.5% | 76.5% |
| Claude Opus 4.7 | 38.5% | 73.6% |
| Claude Opus 4.8 | 46.5% | 74.3% |
| Claude Opus 5 | 53.4% | 78.0% |
| Claude Sonnet 5 | 42.7% | 71.6% |
| Gemini 3.7 Flash | 43.6% | 76.1% |
| GPT-5.4 mini | 27.0% | 56.1% |
| GPT-5.5 | 43.0% | 74.9% |
| GPT-6 Astra | 53.3% | 76.9% |