Microsoft's AesCode-32B learns aesthetics from image generators to code rich posters and dashboards
Powerful_Evening5495 · reddit · 2026-10-15
Microsoft released AesCode-32B, a code generation model built on the idea "good design, written in code." It learns visual aesthetics from reference images produced by image generators, then generates information-rich posters, slides, dashboards, and webpages as editable, verifiable HTML & CSS.
- Weights are open on Hugging Face (with a GGUF quant from mradermacher)
- Code on GitHub and the paper (arXiv 2610.18860, "Aesthetic Code Generation with Decoupled Cross-Modal Rewards") are public
- Demo outputs include a full hospital analytics dashboard and an event poster generated from text prompts
More from Multimodal
- User updates AI-generated photos of a woman, claiming they're 100% synthetic — IndraVahan · 2026-10-15
- A detailed prompt for a 15-second one-take cinematic volcano disaster video — umesh_ai · 2026-10-15
- Physics benchmark scores 8 video models: top Seedance 2.5 only 57.76/100 — jiqizhixin · 2026-10-15
- H3 Turbo Neat LoRA delivers more realistic 4-step renders in ComfyUI for images and video — Ok_Tale7582 · 2026-10-15
- Spooky season: painterly AI Halloween illustrations with film-poster vibes — egeberkina · 2026-10-15
- Classic artworks reimagined as animated pixel art with Gemini 4 — fofrAI · 2026-10-15