# vLLM-Omni adds text-to-image and reference-image editing with multi-GPU VAE tiling

> Original title: From product scenes and game assets to posters and UI mockups—keep the details you want, change the ones you don’t.

- Source: vLLM
- Published: 2026-11-10T15:43:10.000Z
- Bad Influence: https://badinfluence.news/items/wqudocex1y5zier4erms4eul1
- Original: https://x.com/vllm_project/status/2120180471133069484

## Summary

vLLM-Omni supports text-to-image generation and reference-image editing, according to vLLM. The update adds multi-GPU VAE tiling, CUDA Graph replay, and a faster image-response path.

## Post

From product scenes and game assets to posters and UI mockups—keep the details you want, change the ones you don’t.

vLLM-Omni supports text-to-image and reference-image editing, with multi-GPU VAE tiling, CUDA Graph replay and a faster image-response path.