# ARC-AGI-3 → BioBasicsBench

Calibration · https://benchgap.net/calibration/arcagi3/current/to/biobasicsbench/current · data retrieved 2026-11-18

BioBasicsBench is estimated from ARC-AGI-3 with a linear curve fitted on 6 models measured on both: y = 0.4914·x + 0.3673, R² = 0.96, cross-validated error 3.5 pp. It is used for 3 estimates.

| Estimated model | ARC-AGI-3 | BioBasicsBench | Source |
|---|---|---|---|
| [Claude Fable 5.2](https://benchgap.net/model/claude-fable-5-2.md) | 65.7% | 69.0% | estimated ± 3.5 pp, low confidence |
| [Gemini 4 Carbon](https://benchgap.net/model/gemini-4-carbon.md) | 63.9% | 68.1% | estimated ± 3.5 pp, low confidence |
| [Grok 4.20](https://benchgap.net/model/grok-4-20-beta.md) | 0.1% | 36.8% | estimated ± 3.5 pp, low confidence |

## Fitted on

| Model | ARC-AGI-3 | BioBasicsBench |
|---|---|---|
| [GPT-5.4](https://benchgap.net/model/gpt-5-4.md) | 0.2% | 33.2% |
| [GPT-5.5](https://benchgap.net/model/gpt-5-5.md) | 0.4% | 35.4% |
| [GPT-5.6 Sol](https://benchgap.net/model/gpt-5-6-sol.md) | 7.8% | 45.6% |
| [GPT-6.1 Sol](https://benchgap.net/model/gpt-6-1-sol.md) | 52.7% | 62.6% |
| [GPT-6 Astra](https://benchgap.net/model/gpt-6-astra.md) | 62.7% | 67.0% |
| [GPT-6 Luna](https://benchgap.net/model/gpt-6-luna.md) | 0.1% | 37.6% |
