GPT-6 Astra autonomously rebuilds Palace of Fine Arts in Blender overnight
Recoil42 · reddit · 2026-09-04
Sharif Shameem shared a demo of GPT-6 Astra recreating San Francisco's Palace of Fine Arts in Blender: the model autonomously researched hundreds of reference photos, iterated on the scene, rendered intermediate frames and compared them to references, even digging up a Library of Congress scan documenting the columns' dimensions. The human only occasionally steered (sky color, minor clipping); the bulk ran overnight and the finished render awaited him in the morning — another striking early-access datapoint for long-horizon agentic work.
Related event: GPT-6 Astra Autonomously Recreates Palace of Fine Arts in Blender(2 posts)→
More from Models
- Prediction market: param count of the smallest Astra-class reasoning model in 12 months — teortaxesTex · 2026-09-04
- Dev spends $40 on classifier evals to cut costs: 'hard to use AI when you can't afford intelligence' — zeeg · 2026-09-04
- Astra Reasoning Tiers Don't Scale Linearly: xhigh to max Costs ~90% More — weswinder · 2026-09-04
- Satya Nadella announces GPT-6 Astra now available on Azure via Microsoft Foundry — satyanadella · 2026-09-04
- Prediction: The Strongest Lab's Strongest Model Will Be Openly Downloadable by Q3 2027 — teortaxesTex · 2026-09-04
- Kimi K3 Draft Collection: three open-sourced draft models trained with TorchSpec and vLLM on GB200 — AccBalanced · 2026-09-04