BenchmarksReasoning
ARC-AGI-3
The interactive ARC-AGI-3 generalization benchmark: agents must learn novel game environments from scratch (semi-private set).
ReasoningaccuracyHigher is better
What is ARC-AGI-3?
The interactive ARC-AGI-3 generalization benchmark: agents must learn novel game environments from scratch (semi-private set). evals.report tracks reported ARC-AGI-3 scores with the model, source, status, date, and run caveats attached — official leaderboard scores, vendor-reported launches, and clearly labeled community runs.
Top reported ARC-AGI-3 score: GPT-5.6 Sol — 7.78% (accuracy).
| Model | Lab | Score↓ | Source model | Status | Date | |
|---|---|---|---|---|---|---|
| GPT-5.6 Sol | OpenAI | 7.78% | GPT-5.6 Sol | Verified | Jul 9, 2026 | Details |
| Claude Opus 4.8 | Anthropic | 1.52% | Claude Opus 4.8 (High) | Official | May 28, 2026 | Details |
| GPT-5.6 Terra | OpenAI | 0.8% | GPT-5.6 Terra | Verified | Jul 9, 2026 | Details |
| Claude Opus 4.6 | Anthropic | 0.51% | Anthropic Opus 4.6 (Max) | Official | Feb 5, 2026 | Details |
| GPT-5.5 | OpenAI | 0.43% | GPT-5.5 (High) | Official | Apr 23, 2026 | Details |
| Gemini 3.1 Pro Preview | Google DeepMind | 0.42% | Gemini 3.1 Pro (Preview) | Official | Feb 19, 2026 | Details |
| GPT-5.4 | OpenAI | 0.21% | GPT-5.4 (High) | Official | Mar 5, 2026 | Details |
| Claude Opus 4.7 | Anthropic | 0.18% | Opus 4.7 (High) | Official | Apr 16, 2026 | Details |
| GPT-5.6 Luna | OpenAI | 0.18% | GPT-5.6 Luna | Verified | Jul 9, 2026 | Details |
| Grok 4.20 beta reasoning | xAI | 0.09% | Grok 4.20 (Beta Reasoning) | Official | Mar 9, 2026 | Details |
Each row reports the model’s accuracy on ARC-AGI-3. Click a row for the full run context.