Best GPU for Stable Diffusion 3.5 Large: VRAM Requirements
Stable Diffusion 3.5 Large needs about 30.4 GB of GPU memory in BF16 and 7.6 GB at 4 bits. See which GPUs fit it for inference and fine-tuning.
Stable Diffusion 3.5 Large needs about 30.4 GB of VRAM in FP16, 15.2 GB in FP8 and 7.6 GB at 4 bits for inference. In FP16 it fits on one NVIDIA A100 PCIe 40GB; a 4-bit build fits on one NVIDIA GeForce RTX 5080 16GB.
Stable Diffusion 3.5 Large is a 13.8B-parameter image generation model from Stability AI, released in October 2024 and published under the Stability AI Community License. Its weights alone take 27.6 GB in BF16 and 6.9 GB at 4 bits per parameter; the tables below add runtime memory and fine-tuning, then list the GPUs that hold each setup.
Recommended GPUs for Stable Diffusion 3.5 Large
Best for inference
1 GPU × NVIDIA L4 24GB
FP8 or INT8 weights, 15.2 GB.
Best for fine-tuning
1 GPU × NVIDIA A100 PCIe 40GB
LoRA on BF16 weights, 35.9 GB.
Budget pick
1 GPU × NVIDIA GeForce RTX 5080 16GB
Smallest card class that holds a 4-bit build, 7.6 GB.
Stable Diffusion 3.5 Large VRAM requirements by precision
Model weights plus a 10 % margin for activations and runtime buffers, at one request.
| Precision | Weights | KV cache | Total for inference |
|---|---|---|---|
| BF16 / FP16 | 27.6 GB | Not applicable | 30.4 GB |
| FP8 | 13.8 GB | Not applicable | 15.2 GB |
| INT4 | 6.9 GB | Not applicable | 7.6 GB |
Stable Diffusion 3.5 Large VRAM by task
Fine-tuning rows assume a short sequence, batch 1 and gradient checkpointing; longer sequences add activation memory.
| Task | GPU memory |
|---|---|
| QLoRA (4-bit base) | 9 GB |
| LoRA (BF16 base) | 35.9 GB |
| Full fine-tuning (BF16, AdamW) | 242.9 GB |
GPUs that fit Stable Diffusion 3.5 Large
Number of cards whose combined memory holds each setup, assuming the model is split across them with tensor or pipeline parallelism. A dash means more than eight cards.
| GPU | Memory | BF16 inference (30.4 GB) | FP8 inference (15.2 GB) | INT4 inference (7.6 GB) | QLoRA (9 GB) | LoRA (35.9 GB) |
|---|---|---|---|---|---|---|
| NVIDIA GeForce RTX 5080 16GB | 16 GB | 2 GPUs | 1 GPU | 1 GPU | 1 GPU | 4 GPUs |
| NVIDIA GeForce RTX 4090 24GB | 24 GB | 2 GPUs | 1 GPU | 1 GPU | 1 GPU | 2 GPUs |
| NVIDIA L4 24GB | 24 GB | 2 GPUs | 1 GPU | 1 GPU | 1 GPU | 2 GPUs |
| NVIDIA GeForce RTX 5090 32GB | 32 GB | 1 GPU | 1 GPU | 1 GPU | 1 GPU | 2 GPUs |
| NVIDIA Tesla V100 32GB | 32 GB | 1 GPU | 1 GPU | 1 GPU | 1 GPU | 2 GPUs |
| NVIDIA A100 PCIe 40GB | 40 GB | 1 GPU | 1 GPU | 1 GPU | 1 GPU | 1 GPU |
| NVIDIA L40S 48GB | 48 GB | 1 GPU | 1 GPU | 1 GPU | 1 GPU | 1 GPU |
| NVIDIA Quadro RTX 8000 48GB | 48 GB | 1 GPU | 1 GPU | 1 GPU | 1 GPU | 1 GPU |
| NVIDIA RTX 6000 Ada 48GB | 48 GB | 1 GPU | 1 GPU | 1 GPU | 1 GPU | 1 GPU |
| NVIDIA RTX A6000 48GB | 48 GB | 1 GPU | 1 GPU | 1 GPU | 1 GPU | 1 GPU |
| NVIDIA A100 PCIe 80GB | 80 GB | 1 GPU | 1 GPU | 1 GPU | 1 GPU | 1 GPU |
| NVIDIA A100 SXM 80GB | 80 GB | 1 GPU | 1 GPU | 1 GPU | 1 GPU | 1 GPU |
| NVIDIA H100 PCIe 80GB | 80 GB | 1 GPU | 1 GPU | 1 GPU | 1 GPU | 1 GPU |
| NVIDIA H100 SXM 80GB | 80 GB | 1 GPU | 1 GPU | 1 GPU | 1 GPU | 1 GPU |
| NVIDIA RTX PRO 6000 Blackwell 96GB | 96 GB | 1 GPU | 1 GPU | 1 GPU | 1 GPU | 1 GPU |
| NVIDIA H200 141GB | 141 GB | 1 GPU | 1 GPU | 1 GPU | 1 GPU | 1 GPU |
| NVIDIA B200 180GB | 180 GB | 1 GPU | 1 GPU | 1 GPU | 1 GPU | 1 GPU |
| NVIDIA B300 288GB | 288 GB | 1 GPU | 1 GPU | 1 GPU | 1 GPU | 1 GPU |
| NVIDIA GB300 NVL | 288 GB | 1 GPU | 1 GPU | 1 GPU | 1 GPU | 1 GPU |
What is Stable Diffusion 3.5 Large?
Stable Diffusion 3.5 Large is an 8.1-billion-parameter Multimodal Diffusion Transformer (MMDiT) from Stability AI for text-to-image generation at about one megapixel, with improved image quality, typography and understanding of complex prompts.
- Developer
- Stability AI
- Released
- October 2024
- Parameters
- 13.8B
- License
- Stability AI Community License
About 13.8 billion parameters for the whole pipeline: the 8.1B transformer, the CLIP-L (0.12B), OpenCLIP bigG (0.69B) and T5-XXL (4.8B) text encoders and a VAE. Leaving out the T5 encoder lowers memory at some cost to prompt adherence.
Source: Stable Diffusion 3.5 Large model card (as of November 17, 2026).
Frequently asked questions
- How much VRAM does Stable Diffusion 3.5 Large need?
- For inference, Stable Diffusion 3.5 Large needs about 30.4 GB in BF16, 15.2 GB in FP8, 7.6 GB in INT4, including a 10 % runtime margin. Real usage depends on the inference engine and its settings.
- Can Stable Diffusion 3.5 Large run on a single GPU?
- Yes. In BF16 it needs about 30.4 GB, which fits one NVIDIA A100 PCIe 40GB (40 GB).
- How much GPU memory does fine-tuning Stable Diffusion 3.5 Large take?
- QLoRA (4-bit base): about 9 GB; LoRA (BF16 base): about 35.9 GB; Full fine-tuning (BF16, AdamW): about 242.9 GB. Longer training sequences and larger batches add activation memory on top.
- What is the best GPU for Stable Diffusion 3.5 Large?
- Best for inference: 1 GPU × NVIDIA L4 24GB (FP8 or INT8 weights, 15.2 GB). Best for fine-tuning: 1 GPU × NVIDIA A100 PCIe 40GB (LoRA on BF16 weights, 35.9 GB). Budget pick: 1 GPU × NVIDIA GeForce RTX 5080 16GB (Smallest card class that holds a 4-bit build, 7.6 GB). Check live offers before booking, as availability changes.
- Can Stable Diffusion 3.5 Large run on an RTX 4090?
- Yes. In FP8 it needs about 15.2 GB, which fits one RTX 4090 (24 GB). FP16 needs about 30.4 GB, 2 of these cards.
- Can Stable Diffusion 3.5 Large run on an RTX 5090?
- Yes. In FP8 it needs about 15.2 GB, which fits one RTX 5090 (32 GB). FP16 needs about 30.4 GB, which leaves less than 10 % of the card free.
Estimates use the formulas of the LLM VRAM calculator (decimal GB). Real usage depends on the inference engine, quantization format and memory fragmentation: test on the target GPU before a long rental.
Explore the live market
Every offer on the live market comes from a reviewed supplier and is sold by EverXCloud Inc.