// Open-model VRAM table · family
FLUX: local deployment & VRAM
14 open-weight models from Black Forest Labs (35 official repos incl. quantised builds).
official orgs: black-forest-labslatest 2026-10-14memory = estimate
01Image generation (14)
| Model | Params | BF16 | 4-bit | Smallest GPU (4-bit) | Mac (4-bit) | Licence | Released |
|---|---|---|---|---|---|---|---|
| flux-3-image-klein-pup | 9.41B | 21.2 GB | 6.8 GB | 8GB GPU | Mac 16GB | Commercial OK | 2026-10-14 |
| flux-3-image-klein | 5.55B | 12.7 GB | 4.2 GB | 8GB GPU | Mac 16GB | Custom licence – read it | 2026-10-13 |
| FLUX.2-klein-9b-kv | 9.08B | 20.5 GB | 6.6 GB | 8GB GPU | Mac 16GB | Non-commercial | 2026-03-09 |
| FLUX.2-klein-4B | 3.88B | 9.0 GB | 3.1 GB | 8GB GPU | Mac 16GB | Commercial OK | 2026-01-14 |
| FLUX.2-klein-9B | 9.08B | 20.5 GB | 6.6 GB | 8GB GPU | Mac 16GB | Non-commercial | 2026-01-14 |
| FLUX.2-dev | 32.2B | 71.4 GB | 22.0 GB | RTX 5090 32GB | Mac 64GB | Non-commercial | 2025-11-22 |
| FLUX.1-Krea-dev | 11.9B | 26.7 GB | 8.4 GB | 12GB GPU | Mac 16GB | Non-commercial | 2025-07-07 |
| FLUX.1-Kontext-dev | 11.9B | 26.7 GB | 8.4 GB | 12GB GPU | Mac 16GB | Non-commercial | 2025-05-28 |
| FLUX.1-Canny-dev | 11.9B | 26.7 GB | 8.4 GB | 12GB GPU | Mac 16GB | Non-commercial | 2024-11-20 |
| FLUX.1-Depth-dev | 11.9B | 26.7 GB | 8.4 GB | 12GB GPU | Mac 16GB | Non-commercial | 2024-11-20 |
| FLUX.1-Fill-dev | 11.9B | 26.7 GB | 8.4 GB | 12GB GPU | Mac 16GB | Non-commercial | 2024-11-20 |
| FLUX.1-Redux-dev | — | — | — | — | — | Non-commercial | 2024-11-20 |
| FLUX.1-dev | 11.9B | 26.7 GB | 8.4 GB | 12GB GPU | Mac 16GB | Non-commercial | 2024-07-31 |
| FLUX.1-schnell | 11.9B | 26.7 GB | 8.4 GB | 12GB GPU | Mac 16GB | Commercial OK | 2024-07-31 |
02Other families
Qwen · IBM Granite · NVIDIA Nemotron · Gemma · Mistral / Magistral / Devstral · GLM / CogVideoX (Z.ai) · DeepSeek · Ai2 OLMo / Molmo · Ant Ling / Ring / Ming · Liquid LFM · MiniCPM / VoxCPM · Tencent Hunyuan / Hy · Phi · Llama · Meituan LongCat · Jina embeddings & rerankers · Wan · Kimi · Cohere Command / Aya · Xiaomi MiMo · Microsoft VibeVoice / Fara / Harrier · InternLM / Intern-S · StepFun Step · MiniMax · Baidu ERNIE · NVIDIA Parakeet / Canary speech · Stable Diffusion / Stable Audio · Yi (01.AI) · ByteDance Seed · SmolLM / SmolVLM · BGE embeddings & rerankers · LTX Video · FunAudioLLM speech · gpt-oss · poolside Laguna · Whisper · Baichuan · Thinking Machines Inkling · xAI Grok · rednote dots
Totals include 10% + 0.5GB overhead and one 8K-token KV cache for text models. Estimates only. Method →