image model · stable-diffusion · Windows
SD Can I run Stable Diffusion 1.5 on Nvidia GeForce RTX 4090 (24GB)?
Yes. Stable Diffusion 1.5 runs on Nvidia GeForce RTX 4090 (24GB) at fp16 (~3.7 GB of ~23 GB usable).
Runs at fp16 using ~3.7 GB of ~23 GB usable.
- Peak VRAM
- ~3.7 GB
- Usable on device
- ~23 GB
- Device memory
- 24 GB
- Quant
- fp16
How to run it
Use AUTOMATIC1111 or ComfyUI at fp16. It conditions on an image, not a text prompt; the pipeline offloads each stage off the GPU between passes, keeping peak VRAM near the active stage.
- Type
- image (UNET)
- Parameters
- 860M
- Peak VRAM
- ~3.7 GB at fp16
- Resolution
- 512×512
- License
- CreativeML OpenRAIL-M
- Memory
- 24 GB vram
- Usable for weights
- ~23 GB
- Best runtime
- vLLM (Linux) / Ollama (CUDA)
You could also run
Run Stable Diffusion 1.5 on other hardware
FAQ
Can Nvidia GeForce RTX 4090 (24GB) run Stable Diffusion 1.5?
Yes. Stable Diffusion 1.5 runs on Nvidia GeForce RTX 4090 (24GB) at fp16 (~3.7 GB of ~23 GB usable).
How much VRAM does Stable Diffusion 1.5 need?
Nvidia GeForce RTX 4090 (24GB) has room to spare. At fp16 the realistic peak is ~3.7 GB of VRAM, versus ~4 GB with every component kept resident (no offload). With aggressive CPU offload it drops to ~2 GB, much slower.
What do I use to run Stable Diffusion 1.5 locally?
Stable Diffusion 1.5 runs in AUTOMATIC1111 or ComfyUI (among others). It loads as a diffusion checkpoint plus its image encoder and VAE, not a single chat command.
Sources
VRAM figures are sourced peak-usage anchors at the noted quant, validated 2026-06-15. See methodology.