Stability AI's open text-to-image family, free for research and below a revenue threshold.
Specifications
- Provider
- Stability AI
- Type
- Open-source / open-weight
- Modality
- Image generation
- Category
- Image generation
- License
- Stability Community License
- Released
- October 22, 2024
What it was trained for
Open text-to-image generation using a multimodal diffusion transformer architecture.
Best for
- ▸Open and self-hosted image generation
- ▸Creative and concept art
- ▸Custom fine-tuning and LoRA workflows
- ▸Varied output styles
Capabilities
Text-to-image generationImproved prompt adherenceDiverse styles and outputsMultiple sizes for different hardware
Performance & positioning
Improves prompt adherence, typography, and image quality over earlier Stable Diffusion versions; offered in multiple sizes for different hardware budgets.