ARC-AGI-3

The third generation of the ARC-AGI series: instead of static puzzles, the AI is dropped into small interactive game-like environments it has never seen and has to figure out the rules and solve them on its own. Higher is better.

GPT-6 Astra by OpenAI holds the best ARC-AGI-3 result among tracked models, at 99.9% (released Sep 3 2026). Higher scores are better on this benchmark.

1. GPT-6 Astra (OpenAI) — 99.9%; 2. Claude Opus 5 (Anthropic) — 30.2%.

2 tracked models have a published ARC-AGI-3 score. Scores are the figures reported by each lab at that model's release, so this page is a record of results over time rather than a re-run leaderboard.