Claude Opus 5 scores three times higher than the runner-up on ARC-AGI-3

claudeai · x · 2026-07-25

Anthropic says Claude Opus 5 reaches three times the score of the next best model on ARC-AGI-3, an evaluation built around novel problem solving.

The accompanying chart also frames the result by cost, showing Opus 5 as a strong performer at roughly comparable or lower per-task cost than rivals.

Related event: Anthropic Releases Claude Opus 5 with New SOTA Performance at Half Price(25 posts)→

Original post →

More from Models

Models channel →