Gary Marcus: GPT-6 Astra's ARC-AGI-3 success backs symbolic world model hypothesis
GaryMarcus · x · 2026-09-04
Gary Marcus cites ARC Prize's analysis as strong evidence for his long-held hypothesis that symbolic world models are critical: GPT-6 Astra succeeded on ARC-AGI-3 by building dense, compact symbolic world models.
In environment s5i5, Astra recorded level state, hub orientation and mechanism lengths, mapped operations to exact control coordinates, and wrote ordered plans — using on-the-fly algebraic shorthand to preserve the state that mattered and execute precise multi-step plans across turns.
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