← Back to Model Beat
Models·Oct 3·all news from October 3, 2026

AI agents build 3D scenes from photos but have no idea if they got it right

Researchers have introduced a method called LEGO-Anything that converts individual photographs into editable 3D scenes by generating Blender code. While GPT-6 Astra achieved a 53 percent reconstruction accuracy during testing, the study highlights a fundamental limitation in current AI agents. Across all models evaluated, the systems consistently failed to assess the geometric accuracy of their own outputs, suggesting that these agents cannot yet reliably verify the spatial consistency of the scenes they create.

Covered by 1 source

Related stories

ModelsGemini 4 Argon: our next era of frontier intelligenceSep 30 · 15 sourcesModelsSharing AI progress in mathematicsOct 6 · 15 sourcesModelsMistral Large 4 is Europe's trillion-parameter answer to US models that refuse security workOct 6 · 107 sourcesModelsGPT-6 and Intelligent UI for everyoneOct 7 · 6 sources