# Ant Group Open-Sources Unified Image Generation and Editing Model LLaDA-Image

- Category: Flash (快讯)
- Published: 2026-09-08T12:00:00.000Z
- Source date: 2026-09-08
- HTML: https://www.88z.ai/en/news/flash-20260908-473cdffd9e0c1293
- Markdown: https://www.88z.ai/en/news/flash-20260908-473cdffd9e0c1293.md
- Other language: https://www.88z.ai/news/flash-20260908-473cdffd9e0c1293.md

Ant Group's InclusionAI has open-sourced LLaDA-Image, a 6B-parameter text-to-image model. Featuring self-conditioned learning, unified generation and editing, and TwinFlow distillation that reduces steps from 50 to 2–4, it ranks first among open-source models in both Chinese and English on Qwen-Image-Bench, with a total score between GPT Image 1 and Imagen 4.0 Ultra.

## At a glance

- LLaDA-Image is a 6B-parameter open-source model supporting unified generation and editing
- Ranks first among open-source models in Chinese and English on Qwen-Image-Bench, with a total score between GPT Image 1 and Imagen 4.0 Ultra

## The update

InclusionAI, under Ant Group, has officially open-sourced LLaDA-Image, a text-to-image model with 6 billion parameters. The model incorporates self-conditioned learning to unify generation and editing capabilities. It also applies TwinFlow distillation technology, successfully compressing the inference process from 50 steps to just 2–4 steps, significantly improving efficiency. In tests on the authoritative Qwen-Image-Bench, LLaDA-Image ranked first among all open-source models in both Chinese and English scenarios. Its comprehensive score is strong, falling between GPT Image 1 and Imagen 4.0 Ultra, demonstrating leading performance levels.

## Sources

- [蚂蚁开源](https://mp.weixin.qq.com/s/bYm-YTP9higpLzf-vR9Brw) (2026-09-08)

This article is an AI-assisted synthesis by 88Z AI based on the public sources below. Product claims and availability are attributed to their publishers. It does not represent independent testing or first-hand reporting.