← sources

microsoft/AesCode 8B and 32B

source: reddit (reddit:LocalLLaMA)published 2026-10-11 11:17ingested 2026-10-12 02:39

retained text

https://preview.redd.it/r2ledkyuntuh1.png?width=1684&format=png&auto=webp&s=9ed4518b2d0f20f44835ebbfac605581da186a57

AesCode generates information-rich visual artifacts such as slides, posters, and dashboards as HTML/CSS. The output remains **structured, editable, and verifiable**, but the task poses a distinct challenge: code models cannot see how layout, hierarchy, and color come together on the canvas.

Image generators offer the opposite strength. They compose visually compelling pages but often misrender text, numbers, and logical relationships. AesCode uses an image generated from the same prompt as an aesthetic reference while following the prompt for the required content.

Reference input alone does not solve the problem. Off-the-shelf vision-language models may copy hallucinated content or ignore the reference layout. AesCode separates semantic requirements from visual cues through **graph-structured supervision and decoupled cross-modal rewards**.

AesCode-32B starts from **Qwen3-VL-32B-Instruct** and is trained with cold-start SFT followed by GDPO across seven reward channels. It is the larger released checkpoint and leads the aggregate benchmark results.

[https://huggingface.co/microsoft/AesCode-32B](https://huggingface.co/microsoft/AesCode-32B)

[https://huggingface.co/bartowski/AesCode-32B-GGUF](https://huggingface.co/bartowski/AesCode-32B-GGUF)

[https://huggingface.co/microsoft/AesCode-8B](https://huggingface.co/microsoft/AesCode-8B)

[https://huggingface.co/bartowski/AesCode-8B-GGUF](https://huggingface.co/bartowski/AesCode-8B-GGUF)

https://preview.redd.it/7s9j5oooxtuh1.png?width=734&format=png&auto=webp&s=f4c1c8dc912b5fe6f882387da7bb679eae4d767d

document fetched 2026-10-12 00:37 · canonical https://reddit.com/r/LocalLLaMA/comments/1x35o1y/microsoftaescode_8b_and_32b