Qwen3.8-Max rebuilds apps from screenshots and turns 2D floor plans into 3D
Beyond long-horizon coding, Alibaba's flagship recreates software from a screenshot, generates interactive games and educational animations, and converts two-dimensional floor plans into three-dimensional visualisations — multimodality expressed as construction rather than description.
Most multimodal demos end at recognition: the model tells you what it sees. Rebuilding a working application from a picture of one is a different claim — the image becomes a specification, and the output is executable. That collapses the distance between a design artefact and a running program to a single prompt.
The floor-plan conversion is the same capability in another domain. Reading a 2D architectural drawing and producing a coherent 3D interpretation requires inferring unstated structure — wall heights, spatial relationships, conventions a draughtsman assumed. It is closer to domain reasoning than to image processing.
Taken together these point at where multimodal value is settling: not in generating media, which is now abundant, but in translating between representations that professionals currently move between by hand. Screenshot to application, drawing to model, sketch to interactive artefact — each one removes a slow manual conversion from somebody's working week.
Quartz — Alibaba launches Qwen3.8-Max, its largest AI model yet → · Dot Daily Dose — Alibaba unveils Qwen3.8-Max, its largest and most capable flagship model to date →