Stable Diffusion 3.5 Medium
The transformer is small, but it still loads T5-XXL. Offloading the encoder matters more than the model size suggests.
What you actually download
A diffusion pipeline is three separate models. These sizes are read from the real .safetensors files in the Hugging Face repository, not estimated.
Peak VRAM at 1024×1024, one image
Every combination of precision and offload strategy. Lower precision costs image quality; offloading costs speed.
| Precision | Everything in VRAM | Text encoder on CPU | Sequential CPU offload |
|---|---|---|---|
| BF16 / FP16Reference quality. What the repos ship by default. | 19.1 GB | 7.7 GB | 14.0 GB |
| FP8Roughly half the weight memory. Small quality cost, widely used for Flux and SD 3.5. | 10.9 GB | 5.2 GB | 8.3 GB |
| NF4 / INT4Aggressive 4-bit. Visible quality loss, but it puts big models on small cards. | 10.0 GB | 4.2 GB | 8.3 GB |
Which machines can run it
74 of the 75 devices we profile can run Stable Diffusion 3.5 Medium at 1024×1024 in some configuration. Showing the best-quality setup each one supports.
Showing the 30 devices with the most memory. The full hardware matrix covers all 75.
Licence: Stability Community License
With conditions. Stable Diffusion 3.5 Medium is released under Stability Community License. Commercial use is permitted, but the terms carry use restrictions that travel with the model — read them before shipping.
Compare model licences →Stable Diffusion 3.5 Medium: common questions
- How much VRAM does Stable Diffusion 3.5 Medium need?
- At 1024x1024 and batch 1, Stable Diffusion 3.5 Medium peaks at about 19.1GB with everything resident at BF16, and as little as 4.2GB at NF4 + text encoder on CPU. The weights alone are 16.5GB at BF16 — 4.9GB of transformer, 11.4GB of text encoder and 0.17GB of VAE.
- Which GPUs can run Stable Diffusion 3.5 Medium?
- 74 of the 75 devices we profile can run it at 1024x1024 in some configuration. The Apple Mac Studio (M5 Ultra, 512GB, 2026) handles it best, at about 19.1GB. Lower precision and text-encoder offload widen that list considerably.
- Can Stable Diffusion 3.5 Medium be used commercially?
- With conditions. Stable Diffusion 3.5 Medium is released under Stability Community License. Commercial use is permitted, but the terms carry use restrictions that travel with the model — read them before shipping.
- Does offloading the text encoder help on Stable Diffusion 3.5 Medium?
- Yes, substantially. Stable Diffusion 3.5 Medium uses a large language model as its text encoder — 11.4GB at full precision. Encoding happens once per prompt, so moving it to CPU costs a couple of seconds and frees that 11.4GB for the rest of the run, taking peak VRAM at BF16 from 19.1GB down to 7.7GB.
Weight sizes read from the model repository on 2026-09-12. Activation memory is an approximation, not a benchmark — BFWR does not run a hardware test lab.
Other open image models
Guides that go deeper
Which GPUs have the VRAM these models need, and what they cost right now.
- How to Run Stable Diffusion Locally: Hardware, VRAM and SetupCurrent models from SD 1.5 to Qwen-Image, and the VRAM each needs.Read the guide
- Stable Diffusion LoRA: Base Models, Weights and StackingMatching the base model, setting strength, and combining adapters.Read the guide
- How to Increase VRAM: What Actually Works on PC, Laptop and MacIt depends where the memory physically lives.Read the guide