GPT-6.2 Astra scores a straight 100% on ARC-AGI-3 after launch
mhmazur · x · 2026-11-18
After GPT-6.2 Astra officially launched, a user confirmed the model scored a straight 100% on the ARC-AGI-3 benchmark. The ARC-AGI series has long been considered one of the hardest tests of abstract reasoning, making a perfect score highly unusual.
Related event: GPT-6.2 Astra becomes first model to score 100% on ARC-AGI-3(4 posts)→
More from Models
- Pangram AI detector excels at catching the laziest AI-generated text, analyst argues — alejandroll10 · 2026-11-18
- Microsoft's Ignite slide pegs Copilot at $30/month vs ChatGPT $81 and Claude $205 — tomwarren · 2026-11-18
- Microsoft unveils 6 MAI models at Ignite, directly challenging OpenAI's GPT-6.2 — igpila · 2026-11-18
- GPT-6.2 Astra won't launch on Cursor, leaving users unable to test it — IndraVahan · 2026-11-18
- Liquid AI's Maxime Labonne on post-training: SFT, RL, MOPD and regression — ziv_ravid · 2026-11-18
- GPT-6.2 Astra scores 29.01% on Vals RSI-Index, still below Opus 5's ~33% — teortaxesTex · 2026-11-18