AI Trends

ARC Prize announces evaluation results of GPT-6.2 Astra and Sol on ARC-AGI

ARC PrizeUpdated Nov 18 02:30

AI summary

ARC Prize announces that OpenAI GPT-6.2 Astra achieved a SOTA score of 100.0% on ARC-AGI-3, becoming the first model to earn a perfect score on that benchmark, with 95.8% on ARC-AGI-2 and 98.5% on ARC-AGI-1. Sol reached a maximum of 98.2% on ARC-AGI-3, 1.8 percentage points lower than Astra, but at 64% lower cost under the provider adapter approach. ARC Prize also provides links to reproducible results, testing policy, and the complete results for Astra and Sol.

Why read this

SelectedAI score 76

ARC Prize announces the scores and per-task costs of GPT-6.2 Astra and Sol on three versions of ARC-AGI, allowing side-by-side comparison of capability and expense.

Topics

Tags

Body

@OpenAI - Leaderboard: https://t.co/x8U1ItNLRj - Reproduce the public results: https://t.co/vMqtds8Zet - Testing policy: https://t.co/pBoUqvXe4P - Full GPT-6.2 Astra & Sol results: https://t.co/AGc7MkBmkl

Source

ARC Prize