--- format: "aidr-story-markdown/v1" id: "cbbd34e6a2d774b27118afa48221b955edaf3b29c881f716d9cf51714ea8b3cf" canonical_url: "https://aidr.today/cbbd34e6?lang=en" title: "AI agents build 3D scenes from photos but have no idea if they got it right" lang: "en" requested_lang: "en" available_langs: ["en","vi"] translation_fallback: null fallback_fields: [] published_at: "2026-10-03T08:31:04.000Z" category: "Research" topics: ["llm","agent","3d-reconstruction","benchmark","reasoning","safety","weekly-roundup","low-content"] source_urls: ["https://the-decoder.com/ai-agents-build-3d-scenes-from-photos-but-have-no-idea-if-they-got-it-right/","https://www.latent.space/p/ainews-not-much-happened-today-cee"] summary: "A new approach called LEGO-Anything turns single photos into editable Blender code for 3D scenes. GPT-6 Astra leads the accompanying benchmark with up to 53 percent reconstruction accuracy. The biggest weakness across all tested agents is that they can't judge their own geometric accuracy any better than a coin flip. The article AI agents build 3D scenes from photos but have no idea if they got it right appeared first on The Decoder ." --- # AI agents build 3D scenes from photos but have no idea if they got it right > [Open the canonical story]() **Published:** 2026-10-03T08:31:04.000Z **Category:** Research **Topics:** llm, agent, 3d\-reconstruction, benchmark, reasoning, safety, weekly\-roundup, low\-content ## Summary A new approach called LEGO\-Anything turns single photos into editable Blender code for 3D scenes\. GPT\-6 Astra leads the accompanying benchmark with up to 53 percent reconstruction accuracy\. The biggest weakness across all tested agents is that they can't judge their own geometric accuracy any better than a coin flip\. The article AI agents build 3D scenes from photos but have no idea if they got it right appeared first on The Decoder \. ## Sources - [Story source]() - [Story source]()