Baidu · 2026-04-15 · notable
ERNIE-Image — Baidu's Open-Weight Text-to-Image Diffusion Transformer
Baidu open-sources an 8B Diffusion Transformer for text-to-image under Apache 2.0, with a Prompt Enhancer for complex layouts. A fast Turbo variant (8 inference steps) uses DMD and RL distillation.
Baidu's open-weight 8B diffusion transformer for text-to-image, with a Turbo variant that generates in 8 steps.
Key specs
| Parameters | 8B DiT |
|---|---|
| Turbo steps | 8 |
| Geneval | 0.8856 |
| Long text bench | 0.9733 |
What is it?
ERNIE-Image is Baidu's first standalone open-weight text-to-image model. It uses a single-stream Diffusion Transformer (DiT) with 8B parameters and focuses on two known hard problems: accurately rendering dense text inside images (posters, infographics) and following complex multi-object instructions. Released under Apache 2.0 with ComfyUI, Diffusers, and SGLang support.
How does it work?
The DiT uses Flux's VAE for image encoding and Ministral 3.3B as the text encoder. A lightweight Prompt Enhancer expands brief inputs into richer structured descriptions before the DiT runs. A companion Turbo variant distilled via DMD and RL reduces generation to 8 inference steps while preserving quality. Requires a GPU with 24GB VRAM.
Why does it matter?
Text rendering inside generated images is a persistent weak spot for open-weight models. ERNIE-Image scores 0.9733 on LongTextBench and 0.8856 on GENEval, competitive with models several times its size. Apache 2.0 and native ComfyUI/Diffusers support mean it drops directly into existing pipelines.
Who is it for?
Developers building image generation pipelines that need accurate in-image text or complex layout control.
Try it
from diffusers import ErnieImagePipeline