Stable Diffusion XL
The workhorse open-weight image model.
Open Source 6–12 GB VRAM
- Min VRAM
- 6 GB
- GPU class
- Mid GPU
- Quant
- FP16
Actually FreeNo SignupOpen SourceWatermark-Free
FLUX.1 [dev]
12B parameter open-weight diffusion model.
Open Source 8–24 GB VRAMRuns locally
Actually FreeNo SignupOpen SourceWatermark-FreeHobbyist-FriendlyAPIPlugin
Requires: ComfyUI, Stable Diffusion WebUI Forge
Visit FLUX.1 [dev]Updated 2026-05-09 · Direct link
Hardware requirements
Runs locally · High-end GPU (16–24 GB)
Min VRAM
8 GB
Rec. VRAM
24 GB
Min RAM
32 GB
Rec. RAM
64 GB
Disk
30 GB
GPU class
High-end GPU
CUDA 12Apple Silicon ✓GPU RequiredQuant: FP16, FP8, NF4 +1
NF4 fits in 6 GB but loses fidelity. FP8 in 12 GB is the practical sweet spot.
Screenshot placeholder · FLUX.1 [dev]
What is FLUX.1 [dev]?
Black Forest Labs' FLUX.1 [dev] is the current open-weight quality leader for text-to-image. Excellent prompt adherence, hands, and text rendering. Runs in ComfyUI / Forge with FP8 or NF4 quantization for consumer GPUs.
Pros & cons
Pros
- Best open-weight image quality
- Great hands and text
- Quantized variants fit consumer GPUs
Cons
- Non-commercial license on [dev]
- Slow on <12 GB cards even quantized
- No native ControlNet (community ones exist)
What's actually free?
Open weights under a non-commercial license. Run free on your own GPU.
✓ Actually FreeNo SignupOpen SourceWatermark-Free