# ARC-AGI-1 → HLE

Calibration · https://benchgap.net/calibration/arcagi1/current/to/hle/current · data retrieved 2026-11-18

HLE is estimated from ARC-AGI-1 with a offset logistic curve fitted on 16 models measured on both: y = 0.4212 + (0.6476 − 0.4212) / (1 + exp(−60.69·(x − 0.9413))), R² = 0.67, cross-validated error 7.0 pp. It is used for 1 estimate.

| Estimated model | ARC-AGI-1 | HLE | Source |
|---|---|---|---|
| [GPT-5.1](https://benchgap.net/model/gpt-5-1.md) | 72.8% | 42.1% | estimated ± 7.0 pp, medium confidence |

## Fitted on

| Model | ARC-AGI-1 | HLE |
|---|---|---|
| [Claude Fable 5.1](https://benchgap.net/model/claude-fable-5-1.md) | 97.5% | 65.0% |
| [Claude Opus 4.8](https://benchgap.net/model/claude-opus-4-8.md) | 92.5% | 57.9% |
| [Claude Opus 5](https://benchgap.net/model/claude-opus-5.md) | 97.5% | 64.7% |
| [Claude Sonnet 4.6](https://benchgap.net/model/claude-sonnet-4-6.md) | 86.0% | 49.0% |
| [DeepSeek V4 Flash 0731](https://benchgap.net/model/deepseek-v4-flash-0731.md) | 89.0% | 34.8% |
| [DeepSeek V4 Pro](https://benchgap.net/model/deepseek-v4-pro.md) | 90.0% | 42.7% |
| [Gemini 3.5 Flash](https://benchgap.net/model/gemini-3-5-flash.md) | 92.5% | 40.2% |
| [GPT-5.4](https://benchgap.net/model/gpt-5-4.md) | 93.7% | 52.1% |
| [GPT-5.4 mini](https://benchgap.net/model/gpt-5-4-mini.md) | 63.7% | 41.5% |
| [GPT-5.4 nano](https://benchgap.net/model/gpt-5-4-nano.md) | 51.5% | 37.7% |
| [GPT-5.4 Pro](https://benchgap.net/model/gpt-5-4-pro.md) | 94.5% | 58.7% |
| [GPT-5.5](https://benchgap.net/model/gpt-5-5.md) | 95.0% | 52.2% |
| [GPT-5.5 Pro](https://benchgap.net/model/gpt-5-5-pro.md) | 95.0% | 57.2% |
| [Inkling 2 Small](https://benchgap.net/model/inkling-2-small.md) | 96.5% | 53.2% |
| [Inkling-Small](https://benchgap.net/model/inkling-small.md) | 84.0% | 47.8% |
| [Kimi K3](https://benchgap.net/model/kimi-k3.md) | 94.5% | 56.0% |
