black-forest-labs

FLUX.1 [dev]

Black Forest Labs' 12B text-to-image model, one of the most-liked open image models on Hugging Face.text to imageotherNo commercial useAccess request needed495.6K downloadsHugging Face

Cheapest GPU in stock that fits

A single GPU with at least 24 GB for inference.

NosanaIn stock · 9 freeStart right away

1× RTX 309024 GB VRAM

Nosana network, hosts worldwide

Details

Stock and price checked 15 min ago.

Self-service: you start it yourself, with no order or sales contact.

nvidia-3090

Full price

€0.17

per hour, whole instance, 1 GPU

View offer

License restrictions

Weights are for non-commercial use only without a separate BFL license; the license also addresses outputs, so read it before commercial use of images. Repo is gated.

GPU memory it needs

From model cards and vendor documentation. Lower precision needs less memory, with slightly lower quality. Estimates are marked.

What you want to do	Precision	Minimum	Recommended	Setup	Basis
Run the model NVIDIA states 'over 23GB' at FP16. Its table for the FLUX pipeline shows 23.9 GB in low-VRAM mode (minimum here) and 39.3 GB in default mode (recommended here); the table names only 'FLUX'.	FP16, full	24 GB	39.3 GB	—	Published
Run the model NVIDIA table for the FLUX pipeline: 14.9 GB in low-VRAM mode, 24.6 GB default.	FP8, 8-bit	14.9 GB	24.6 GB	—	Published
Run the model NVIDIA table for FP4: 11.1 GB in low-VRAM mode, 21.67 GB default. NVIDIA's RTX 50 post states under 10 GB for FP4 and over 23 GB for FP16.	4-bit	11.1 GB	21.67 GB	GeForce RTX 50 series for FP4, per NVIDIA	Published