SD 3.5 Large
imageby Stability AI · Updated October 1, 2025
Stable Diffusion 3.5 Large is Stability AI's most capable open-source image model with 8 billion parameters. It delivers significant improvements in prompt adherence, text rendering, fine details, and overall coherence compared to SD3 Medium. Available under a permissive license for both personal and commercial use.
Best For
Prompting Tips
- 1Handles complex multi-subject compositions better than previous SD versions
- 2Text rendering is significantly improved — specify font styles
- 3Use CFG scale 4-7 for best results
- 4Supports LoRA and ControlNet fine-tuning
- 5Natural language prompts work well — less keyword stacking needed than SDXL
Syntax & Constraints
Natural language prompts with optional weighting. Negative prompt supported. CFG scale 4-7 recommended. 1024x1024 native resolution with other ratios supported.
Example Prompts
👤 Portrait Photography
Architect environmental portrait
Professional environmental portrait of a female architect in her modern glass-walled office, city skyline visible behind her. Confident pose, arms crossed, wearing a tailored navy suit. Natural daylight, shallow depth of field on her face. Editorial magazine quality.
Build Prompts for SD 3.5 Large
Visual prompt builder
Visual prompt builder
Visual prompt builder
Visual prompt builder
Visual prompt builder
Visual prompt builder
Visual prompt builder
Visual prompt builder
Visual prompt builder
Visual prompt builder
Visual prompt builder
Visual prompt builder
Visual prompt builder
Visual prompt builder
Visual prompt builder
Visual prompt builder
Visual prompt builder
Visual prompt builder
Visual prompt builder
Visual prompt builder
Visual prompt builder
Visual prompt builder
Visual prompt builder
Visual prompt builder
Other Stable Diffusion Models
The most popular open-source image generation model.
Next-gen architecture with improved text and composition.
Efficient multi-stage generation architecture.
Real-time SDXL generation in a single step.
Open-source image-to-video generation model.
Stability AI's open-source music and sound generation model with 3-minute high-quality output.