Image models · BAAI (VectorSpaceLab)
OmniGen2
- Backbone
- 4.0B
- Native output
- 1024×1024
- Default steps
- 50
- License
- apache-2.0
- Released
- 2025-06
How to run it
Original weights · Hugging Face
Full-precision files from BAAI (VectorSpaceLab), for fine-tuning, vLLM, Transformers, Diffusers or ComfyUI.
Official repo · GitHub
Reference code, inference scripts and docs from BAAI (VectorSpaceLab). The weights themselves download from Hugging Face.
github.com/VectorSpaceLab/OmniGen2 (opens in a new tab)
git clone https://github.com/VectorSpaceLab/OmniGen2Memory needed
Peak GPU memory at the native output size. Offloading moves the text encoder, or the whole pipeline, to system RAM between steps.
| Precision | Everything on GPU | Text encoder offloaded | Sequential offload | Download |
|---|---|---|---|---|
FP16 | 18.0 GB | 10.5 GB | 5.2 GB | Original weights (opens in a new tab) |
FP8 | 10.3 GB | 6.6 GB | 3.8 GB | Original weights (opens in a new tab) |
Components
- Backbone: 3.97B
- VAE: 0.084B
- Qwen2.5-VL 3B: 3.75B
Best hardware for OmniGen2
The least expensive GPU in our catalog that runs it without offloading is the Intel Arc A380 (launch price $139) at FP8.
| Grade · points | Hardware | Precision | Memory | Time per output | Fit |
|---|---|---|---|---|---|
| 79 ptsGreat (A) · 79 of 100 points: Model quality 29.1/45, Memory fit 25.0/25, Speed 10.9/15, Memory use 14.9/15 | AMD Radeon RX 9070 XT | FP8 | 10.3 GB / 16.0 GB | 19 s | Runs well |
| 79 ptsGreat (A) · 79 of 100 points: Model quality 29.1/45, Memory fit 25.0/25, Speed 10.8/15, Memory use 14.9/15 | NVIDIA GeForce RTX 5080 | FP8 | 10.3 GB / 16.0 GB | 19 s | Runs well |
| 79 ptsGreat (A) · 79 of 100 points: Model quality 29.1/45, Memory fit 25.0/25, Speed 10.5/15, Memory use 14.9/15 | NVIDIA GeForce RTX 4080 SUPER | FP8 | 10.3 GB / 16.0 GB | 20 s | Runs well |
| 78 ptsGreat (A) · 78 of 100 points: Model quality 29.1/45, Memory fit 25.0/25, Speed 9.9/15, Memory use 14.9/15 | NVIDIA GeForce RTX 4070 Ti SUPER | FP8 | 10.3 GB / 16.0 GB | 23 s | Runs well |
| 78 ptsGreat (A) · 78 of 100 points: Model quality 29.1/45, Memory fit 25.0/25, Speed 9.9/15, Memory use 14.9/15 | NVIDIA GeForce RTX 5070 Ti | FP8 | 10.3 GB / 16.0 GB | 23 s | Runs well |
| 78 ptsGreat (A) · 78 of 100 points: Model quality 29.7/45, Memory fit 25.0/25, Speed 8.6/15, Memory use 15.0/15 | NVIDIA GeForce RTX 3090 | FP16 | 18.0 GB / 24.0 GB | 30 s | Runs well |
| 76 ptsGreat (A) · 76 of 100 points: Model quality 29.1/45, Memory fit 25.0/25, Speed 7.8/15, Memory use 14.9/15 | NVIDIA GeForce RTX 5060 Ti 16GB | FP8 | 10.3 GB / 16.0 GB | 37 s | Runs well |
| 76 ptsGreat (A) · 76 of 100 points: Model quality 29.1/45, Memory fit 25.0/25, Speed 7.6/15, Memory use 14.9/15 | NVIDIA GeForce RTX 4060 Ti 16GB | FP8 | 10.3 GB / 16.0 GB | 39 s | Runs well |
| 76 ptsGreat (A) · 76 of 100 points: Model quality 29.1/45, Memory fit 25.0/25, Speed 12.5/15, Memory use 9.9/15 | NVIDIA GeForce RTX 4090 | FP8 | 10.3 GB / 24.0 GB | 14 s | Runs well |
| 76 ptsGreat (A) · 76 of 100 points: Model quality 29.1/45, Memory fit 23.0/25, Speed 9.1/15, Memory use 15.0/15 | NVIDIA GeForce RTX 4070 SUPER | FP8 | 10.3 GB / 12.0 GB | 27 s | Tight fit |
Estimates come from published hardware specs and our open formulas, not from running the model on your machine. Real speed varies with runtime, drivers and settings. Numbers marked “Measured” come from a linked benchmark.