We're excited to announce the official launch of Z Image Base, a stable and versatile AI image generation foundation model:
- Text to Image: Generate high-quality images from natural language descriptions in Chinese or English
- Image to Image: Transform images while preserving structure and composition
- Multiple Resolutions: Support for 1024×1024, 2048×2048, and more
- Batch Generation: Generate up to 4 images at once
- History Management: Save and review your generation history
- 6B Parameters: Balanced quality with accessible deployment requirements
- S3-DiT Architecture: Single-stream Diffusion Transformer for efficiency and quality
- Structural Stability: Consistent human proportions and object structures
- Bilingual Support: Excellent understanding of Chinese and English prompts
- Apache 2.0 License: Free for commercial use
- 16GB VRAM: Runs on consumer-grade GPUs
- LoRA Friendly: Complete undistilled model, ideal for fine-tuning