A new technology called LEGO-Anything transforms single photographs into editable 3D scenes by generating Blender code, according to The Decoder. This approach aims to simplify 3D content creation by enabling users to convert flat images directly into manipulable 3D environments.
The latest benchmark results show that GPT-6 Astra achieves the highest reconstruction accuracy, reaching up to 53 percent. However, The Decoder notes that all tested AI agents struggle to reliably assess their own geometric accuracy, performing no better than a coin toss in judging the quality of their 3D reconstructions.
For Japanese markets, where 3D visualization and virtual content are increasingly integrated into gaming and digital marketing, advancements like LEGO-Anything could accelerate creative workflows and enhance interactive experiences across sectors including FX and crypto platforms.