François Chollet just released ARC-AGI-3 — the hardest AI test ever created.

Humans: 100%
Gemini 3.1 Pro: 0.37%
GPT 5.4: 0.26%
Opus 4.6: 0.25%
Grok-4.20: 0.00%

25 games are open so you can try them and see the difference.

Post to X