ARC Prize measured OpenAI GPT-6 Astra on ARC-AGI-3 Semi-Private: about 62.7% for ~$26k with the Standard harness and 99.9% for ~$19k with a Provider Adapter harness that preserves opaque reasoning state. Astra beat the human action-efficiency baseline on 96% of levels and built custom algebraic notes and tools.
Key Takeaways
- ✓Standard harness: ~62.7% / ~$26k on ARC-AGI-3 Semi-Private under a provider-neutral interface.
- ✓Provider Adapter: keeps opaque reasoning state and compaction — ~99.9%, ~49% fewer tokens.
- ✓Action efficiency: fewer actions than the human median on 96% of levels; builds custom notation and game-specific tools.
Discussion & Comments
0Sign in to join the discussion
Connect with AI developers to exchange benchmark insights.