AI agents build 3D scenes from photos but have no idea if they got it right
A new approach called LEGO-Anything turns a single photo into editable Blender code for 3D scenes, according to The Decoder. The method is accompanied by a benchmark that evaluates AI agents on this photo-to-3D task.
GPT-6 Astra leads that benchmark with up to 53 percent reconstruction accuracy.
The reported weakness across all tested agents is self-assessment: they cannot judge their own geometric accuracy any better than a coin flip.
The benchmark results and the self-evaluation finding apply to the agents evaluated under the LEGO-Anything work. The available information does not detail the benchmark's size, the number of agents tested, or their individual scores beyond the top result.
Based on reporting from the original publisher. Visit the source for full context and later updates.
Publisher excerpt
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 .