Calibration
Vals GPQA Diamond → ARC-AGI-1
ARC-AGI-1 is estimated from Vals GPQA Diamond with a Hill curve fitted on 21 models measured on both: y = 0.0000 + (1.2000 − 0.0000)·x^6.00 / (0.76014^6.00 + x^6.00), R² = 0.76, cross-validated error 5.9 pp. It is used for 14 estimates.
| Estimated model | Vals GPQA Diamond | ARC-AGI-1 | Source |
|---|---|---|---|
| MiniMax M3 | 92.7% | 92.0% | estimated ± 5.9 pp, medium confidence |
| Grok 4.3 | 91.4% | 90.2% | estimated ± 5.9 pp, medium confidence |
| GLM-5.3 | 88.1% | 85.0% | estimated ± 5.9 pp, medium confidence |
| Gemini 3 Flash | 87.9% | 84.6% | estimated ± 5.9 pp, medium confidence |
| Qwen3.6 Plus | 87.4% | 83.8% | estimated ± 5.9 pp, medium confidence |
| GLM-5.3-Flash | 86.4% | 82.0% | estimated ± 5.9 pp, medium confidence |
| Gemini 3.5 Flash-Lite | 83.8% | 77.1% | estimated ± 5.9 pp, medium confidence |
| Qwen3.5 Flash | 82.8% | 75.1% | estimated ± 5.9 pp, medium confidence |
| MiMo-V2.5 | 81.6% | 72.6% | estimated ± 5.9 pp, medium confidence |
| Gemini 3.1 Flash-Lite | 81.1% | 71.5% | estimated ± 5.9 pp, medium confidence |
| GLM-4.7 | 80.1% | 69.3% | estimated ± 5.9 pp, medium confidence |
| GLM-4.6 | 74.5% | 56.4% | estimated ± 5.9 pp, low confidence |
| Claude Haiku 4.5 | 72.2% | 50.8% | estimated ± 5.9 pp, low confidence |
| GLM-4.5 | 72.2% | 50.8% | estimated ± 5.9 pp, low confidence |
Fitted on
| Model | Vals GPQA Diamond | ARC-AGI-1 |
|---|---|---|
| Claude Fable 5 | 93.2% | 98.5% |
| Claude Fable 5.1 | 93.4% | 97.5% |
| Claude Opus 4.8 | 92.4% | 92.5% |
| Claude Opus 5 | 93.4% | 97.5% |
| Claude Sonnet 4.6 | 85.6% | 86.0% |
| DeepSeek V4 Flash 0731 | 89.9% | 89.0% |
| DeepSeek V4 Pro 0813 | 92.4% | 90.0% |
| Gemini 3.5 Flash | 92.7% | 92.5% |
| Gemini 3.6 Flash | 93.4% | 91.2% |
| Gemini 3.7 Flash | 93.9% | 95.5% |
| Gemini 3.8 Flash | 94.4% | 98.5% |
| GPT-5.4 mini | 83.1% | 63.7% |
| GPT-5.4 nano | 77.5% | 51.5% |
| GPT-5.5 | 93.2% | 95.0% |
| GPT-5.6 Luna | 91.7% | 88.0% |
| GPT-5.6 Sol | 95.2% | 96.5% |
| GPT-5.6 Terra | 90.9% | 96.5% |
| Grok 4.5 | 92.9% | 85.7% |
| Grok 4.6 | 94.7% | 87.0% |
| Inkling-Small | 83.6% | 84.0% |
| Kimi K3 | 92.9% | 94.5% |