Home › Prompts › Stable Diffusion › SD 3.5 Large photoreal with negative prompt

SD 3.5 Large photoreal with negative prompt

A local or API photoreal render using SD 3.5's natural-language prompts, negative prompt and sampler settings.

Stable DiffusionImage generationRuleStack original

Written for Stable Diffusion 3.5 Large · Updated 2026-10-16

Note: Review and test before use. Provided as is, without warranty. See Terms.

Positive: A [shot type] photograph of [subject] in [location], [lighting], [lens and aperture], natural colors, fine detail, realistic textures.
Negative: blurry, oversaturated, extra fingers, distorted hands, watermark, text, jpeg artifacts

Replace everything in [square brackets] with your own details.

Settings & parameters

• Model: SD 3.5 Large (8B); Large Turbo for 4-step drafts; Medium (2.5B) for consumer GPUs
• Steps ~28–40, CFG ~3.5–4.5 for Large (Turbo: 4 steps, CFG ~1)
• Resolution around 1 megapixel (e.g. 1024x1024, 1152x896)
• Runs in ComfyUI or diffusers; check the Stability Community License for commercial use

Variations

• Use Large Turbo to explore, then rerun the best seed on Large.
• Add a ControlNet (depth/canny) to lock composition.
• Swap to an illustration description; SD 3.5 handles many styles.

Tips

• SD 3.5 understands sentences; you don't need weighted keyword soup.
• Keep the negative prompt short and specific.