OpenAI's Astra Tops ARC-AGI-3 Benchmark

OpenAI's new model Astra scored 99.9% on ARC-AGI-3, far surpassing prior models at around 30%. Even with reasoning disabled it reached 35.2%, over 4.5 times higher than GPT-5.6.

2026-09-09 ~ 2026-09-09 · 2 related posts

• GPT-6 Astra scores 35.2% on ARC-AGI-3 with reasoning set to 'none', 4.5x GPT-5.6's max — mhmazur · 2026-09-09
• OpenAI's Astra scores 99.9% on ARC-AGI-3, far surpassing prior models' ~30% — downingARK · 2026-09-09