GPT-6 Astra saturates ARC-AGI-3 using fewer moves than average humans
TimeTruth2490 · reddit · 2026-09-04
According to the official ARC Prize blog, GPT-6 Astra not only saturates ARC-AGI-3 but completes tasks using fewer moves than the average human. The result suggests ARC-AGI-3's discriminating power as a general reasoning benchmark may have been broken by frontier models.
Related event: GPT-6 Astra Reportedly Saturates ARC-AGI-3 with Fewer Steps Than Humans(5 posts)→
More from Models
- Scaling Law for Looped Transformers: Looping Boosts Reasoning, Not Knowledge — bookwormengr · 2026-09-04
- AI experts once pegged AGI at 2075-2100 — OpenAI's Astra shows how wrong they were — dee_hw · 2026-09-04
- Models still can't see determinants, but no-thinking time-horizons double every ~9 months — scaling01 · 2026-09-04
- GPT-6 Astra Ultra builds a Minecraft wooden house zero-shot in ~9 minutes, stairs perfect — Angaisb_ · 2026-09-04
- Benchmark saturation accelerates: ARC-AGI-3 falls in 5 months after ARC-AGI-1's 6 years — mmmbchang · 2026-09-04
- Sam Altman apologizes for messy Astra rollout, promises banked credit resets and broad API rollout soon — jxnlco · 2026-09-04