FLUX.2 [klein] 9B VRAM requirements
A distilled 9B FLUX.2 for fast generation and editing in a few steps. Much lighter than FLUX.2 dev.
TypeImage
Parameters9B
Fits entirely from7 GB
8-bit or better from12 GB
01Measured on my RTX 5060 Ti 16 GB
Q8_0 · 1024×1024FASTEST
8.5s / image
flux-2-klein-9b-Q8_0.gguf
Peak VRAM13.5 GB
Run range8.5–8.6 s
Q4_K_M · 1024×1024
9.8s / image
flux-2-klein-9b-Q4_K_M.gguf
Peak VRAM13.5 GB
Run range9.8–9.9 s
02The files, and how much VRAM each needs
| File | Size | Needed | Min. VRAM | Quality | Source |
|---|---|---|---|---|---|
| 16-bit | 18.2 GB | 20.5 GB | 21 GB | the original weights | black-forest-labs/FLUX.2-klein-9B → |
| Q8_0 | 10.0 GB | 12.3 GB | 13 GB | practically identical to the original | unsloth/FLUX.2-klein-9B-GGUF → |
| FP8 | 9.4 GB | 11.7 GB | 12 GB | practically identical to the original | black-forest-labs/FLUX.2-klein-9b-fp8 → |
| Q6_K | 7.9 GB | 10.2 GB | 11 GB | very close to the original | unsloth/FLUX.2-klein-9B-GGUF → |
| Q5_K_M | 7.0 GB | 9.3 GB | 10 GB | close; small differences in fine detail | unsloth/FLUX.2-klein-9B-GGUF → |
| Q4_K_M | 5.9 GB | 8.2 GB | 9 GB | good; some loss in fine detail and text | unsloth/FLUX.2-klein-9B-GGUF → |
| Q3_K_M | 4.8 GB | 7.1 GB | 8 GB | noticeable loss of detail | unsloth/FLUX.2-klein-9B-GGUF → |
| Q2_K | 4.0 GB | 6.3 GB | 7 GB | heavy loss; a last resort | unsloth/FLUX.2-klein-9B-GGUF → |
“Needed” = file + 1.5 GB working memory + 0.8 GB system reserve.
Free ComfyUI workflow, tested on my RTX 5060 Ti: GGUF Q8_0 workflow (.json). Drag the file into ComfyUI. All workflows →
03By amount of VRAM
04Best GPU for FLUX.2 klein 9B
The cheapest cards (by launch price) that run it well, and every card sorted by memory: best GPU for FLUX.2 klein 9B → Planning bigger images or longer clips? Open the calculator →
05By graphics card
| GPU | VRAM | Verdict | Best file | Needed |
|---|---|---|---|---|
| Desktop graphics cards | ||||
| RTX 2060 6 GB | 6 GB | Offload only | Q3_K_M | 7.1 GB |
| RTX 3050 6 GB | 6 GB | Offload only | Q3_K_M | 7.1 GB |
| RTX 2070 Super 8 GB | 8 GB | Tight | Q3_K_M | 7.1 GB |
| RTX 2080 Super 8 GB | 8 GB | Tight | Q3_K_M | 7.1 GB |
| RTX 3050 8 GB | 8 GB | Tight | Q3_K_M | 7.1 GB |
| RTX 3060 8 GB | 8 GB | Tight | Q3_K_M | 7.1 GB |
| RTX 3060 Ti 8 GB | 8 GB | Tight | Q3_K_M | 7.1 GB |
| RTX 3070 8 GB | 8 GB | Tight | Q3_K_M | 7.1 GB |
| RTX 3070 Ti 8 GB | 8 GB | Tight | Q3_K_M | 7.1 GB |
| RTX 4060 8 GB | 8 GB | Tight | Q3_K_M | 7.1 GB |
| RTX 4060 Ti 8 GB | 8 GB | Tight | Q3_K_M | 7.1 GB |
| RTX 5050 8 GB | 8 GB | Tight | Q3_K_M | 7.1 GB |
| RTX 5060 8 GB | 8 GB | Tight | Q3_K_M | 7.1 GB |
| RTX 5060 Ti 8 GB | 8 GB | Tight | Q3_K_M | 7.1 GB |
| RX 7600 8 GB | 8 GB | Tight | Q3_K_M | 7.1 GB |
| RX 9050 8 GB | 8 GB | Tight | Q3_K_M | 7.1 GB |
| RX 9060 XT 8 GB | 8 GB | Tight | Q3_K_M | 7.1 GB |
| Arc B570 10 GB | 10 GB | Runs | Q5_K_M | 9.3 GB |
| RTX 3080 10 GB | 10 GB | Runs | Q5_K_M | 9.3 GB |
| RTX 2080 Ti 11 GB | 11 GB | Runs | Q6_K | 10.2 GB |
| Arc B580 12 GB | 12 GB | Runs well | FP8 | 11.7 GB |
| RTX 2060 12 GB | 12 GB | Runs well | FP8 | 11.7 GB |
| RTX 3060 12 GB | 12 GB | Runs well | FP8 | 11.7 GB |
| RTX 3080 12 GB | 12 GB | Runs well | FP8 | 11.7 GB |
| RTX 3080 Ti 12 GB | 12 GB | Runs well | FP8 | 11.7 GB |
| RTX 4070 12 GB | 12 GB | Runs well | FP8 | 11.7 GB |
| RTX 4070 Super 12 GB | 12 GB | Runs well | FP8 | 11.7 GB |
| RTX 4070 Ti 12 GB | 12 GB | Runs well | FP8 | 11.7 GB |
| RTX 5070 12 GB | 12 GB | Runs well | FP8 | 11.7 GB |
| RX 7700 XT 12 GB | 12 GB | Runs well | FP8 | 11.7 GB |
| RX 9070 GRE 12 GB | 12 GB | Runs well | FP8 | 11.7 GB |
| Arc A770 16 GB | 16 GB | Runs well | Q8_0 | 12.3 GB |
| RTX 4060 Ti 16 GB | 16 GB | Runs well | FP8 | 11.7 GB |
| RTX 4070 Ti Super 16 GB | 16 GB | Runs well | FP8 | 11.7 GB |
| RTX 4080 16 GB | 16 GB | Runs well | FP8 | 11.7 GB |
| RTX 4080 Super 16 GB | 16 GB | Runs well | FP8 | 11.7 GB |
| RTX 5060 Ti 16 GB | 16 GB | Runs well | FP8 | 11.7 GB |
| RTX 5070 Ti 16 GB | 16 GB | Runs well | FP8 | 11.7 GB |
| RTX 5080 16 GB | 16 GB | Runs well | FP8 | 11.7 GB |
| RX 7600 XT 16 GB | 16 GB | Runs well | Q8_0 | 12.3 GB |
| RX 7800 XT 16 GB | 16 GB | Runs well | Q8_0 | 12.3 GB |
| RX 7900 GRE 16 GB | 16 GB | Runs well | Q8_0 | 12.3 GB |
| RX 9060 XT 16 GB | 16 GB | Runs well | FP8 | 11.7 GB |
| RX 9070 16 GB | 16 GB | Runs well | FP8 | 11.7 GB |
| RX 9070 XT 16 GB | 16 GB | Runs well | FP8 | 11.7 GB |
| RX 7900 XT 20 GB | 20 GB | Runs well | Q8_0 | 12.3 GB |
| Arc Pro B60 24 GB | 24 GB | Runs well | 16-bit | 20.5 GB |
| RTX 3090 24 GB | 24 GB | Runs well | 16-bit | 20.5 GB |
| RTX 3090 Ti 24 GB | 24 GB | Runs well | 16-bit | 20.5 GB |
| RTX 4090 24 GB | 24 GB | Runs well | 16-bit | 20.5 GB |
| RX 7900 XTX 24 GB | 24 GB | Runs well | 16-bit | 20.5 GB |
| Arc Pro B70 32 GB | 32 GB | Runs well | 16-bit | 20.5 GB |
| RTX 5090 32 GB | 32 GB | Runs well | 16-bit | 20.5 GB |
| Laptop GPUs | ||||
| RTX 3050 Laptop 4 GB | 4 GB | Offload only | Q4_K_M | 8.2 GB |
| RTX 3050 Ti Laptop 4 GB | 4 GB | Offload only | Q4_K_M | 8.2 GB |
| RTX 2060 Laptop 6 GB | 6 GB | Offload only | Q3_K_M | 7.1 GB |
| RTX 3050 Laptop 6 GB | 6 GB | Offload only | Q3_K_M | 7.1 GB |
| RTX 3060 Laptop 6 GB | 6 GB | Offload only | Q3_K_M | 7.1 GB |
| RTX 4050 Laptop 6 GB | 6 GB | Offload only | Q3_K_M | 7.1 GB |
| RTX 2070 Laptop 8 GB | 8 GB | Tight | Q3_K_M | 7.1 GB |
| RTX 2070 Super Laptop 8 GB | 8 GB | Tight | Q3_K_M | 7.1 GB |
| RTX 2080 Laptop 8 GB | 8 GB | Tight | Q3_K_M | 7.1 GB |
| RTX 2080 Super Laptop 8 GB | 8 GB | Tight | Q3_K_M | 7.1 GB |
| RTX 3070 Laptop 8 GB | 8 GB | Tight | Q3_K_M | 7.1 GB |
| RTX 3070 Ti Laptop 8 GB | 8 GB | Tight | Q3_K_M | 7.1 GB |
| RTX 3080 Laptop 8 GB | 8 GB | Tight | Q3_K_M | 7.1 GB |
| RTX 4060 Laptop 8 GB | 8 GB | Tight | Q3_K_M | 7.1 GB |
| RTX 4070 Laptop 8 GB | 8 GB | Tight | Q3_K_M | 7.1 GB |
| RTX 5050 Laptop 8 GB | 8 GB | Tight | Q3_K_M | 7.1 GB |
| RTX 5060 Laptop 8 GB | 8 GB | Tight | Q3_K_M | 7.1 GB |
| RTX 5070 Laptop 8 GB | 8 GB | Tight | Q3_K_M | 7.1 GB |
| RX 7600M 8 GB | 8 GB | Tight | Q3_K_M | 7.1 GB |
| RX 7600M XT 8 GB | 8 GB | Tight | Q3_K_M | 7.1 GB |
| RX 7600S 8 GB | 8 GB | Tight | Q3_K_M | 7.1 GB |
| RX 7700S 8 GB | 8 GB | Tight | Q3_K_M | 7.1 GB |
| RTX 4080 Laptop 12 GB | 12 GB | Runs well | FP8 | 11.7 GB |
| RTX 5070 Laptop 12 GB | 12 GB | Runs well | FP8 | 11.7 GB |
| RTX 5070 Ti Laptop 12 GB | 12 GB | Runs well | FP8 | 11.7 GB |
| RX 7800M 12 GB | 12 GB | Runs well | FP8 | 11.7 GB |
| RTX 3080 Laptop 16 GB | 16 GB | Runs well | Q8_0 | 12.3 GB |
| RTX 3080 Ti Laptop 16 GB | 16 GB | Runs well | Q8_0 | 12.3 GB |
| RTX 4090 Laptop 16 GB | 16 GB | Runs well | FP8 | 11.7 GB |
| RTX 5080 Laptop 16 GB | 16 GB | Runs well | FP8 | 11.7 GB |
| RX 7900M 16 GB | 16 GB | Runs well | Q8_0 | 12.3 GB |
| RTX 5090 Laptop 24 GB | 24 GB | Runs well | 16-bit | 20.5 GB |
| Unified memory | ||||
| Radeon 8060S (Strix Halo) 96 GB | 96 GB | Runs well | 16-bit | 20.5 GB |
| Radeon 8065S (Gorgon Halo) 160 GB | 160 GB | Runs well | 16-bit | 20.5 GB |
| Apple Silicon Macs (by memory) | ||||
| Mac 16 GB | 12.7 GB | Runs well | Q8_0 | 12.3 GB |
| Mac 18 GB | 14.4 GB | Runs well | Q8_0 | 12.3 GB |
| Mac 24 GB | 19.6 GB | Runs well | Q8_0 | 12.3 GB |
| Mac 32 GB | 26.8 GB | Runs well | 16-bit | 20.5 GB |
| Mac 36 GB | 30.2 GB | Runs well | 16-bit | 20.5 GB |
| Mac 48 GB | 40.2 GB | Runs well | 16-bit | 20.5 GB |
| Mac 64 GB | 55.7 GB | Runs well | 16-bit | 20.5 GB |
| Mac 96 GB | 85 GB | Runs well | 16-bit | 20.5 GB |
| Mac 128 GB | 115.4 GB | Runs well | 16-bit | 20.5 GB |
| Mac 192 GB | 175.4 GB | Runs well | 16-bit | 20.5 GB |
| Mac 256 GB | 236.9 GB | Runs well | 16-bit | 20.5 GB |
| Mac 512 GB | 498.1 GB | Runs well | 16-bit | 20.5 GB |
06Text encoder, VAE and other files
| File | Folder | Size | When | |
|---|---|---|---|---|
| Qwen3 8B FP8 qwen_3_8b_fp8mixed.safetensors | models/text_encoders | 8.7 GB | text encoder · smaller, recommended | Download → |
| Qwen3 8B BF16 qwen_3_8b.safetensors | models/text_encoders | 16.4 GB | text encoder · alternative | Download → |
| FLUX.2 VAE flux2-vae.safetensors | models/vae | 0.3 GB | required | Download → |
The files the official ComfyUI workflows load next to the model. Sizes read from Hugging Face (2026-10-01).
07Where the files go in ComfyUI
| File | Folder | Loader node |
|---|---|---|
| Diffusion model (.safetensors: 16-bit, FP8, INT8) | ComfyUI/models/diffusion_models | Load Diffusion Model |
| GGUF file (.gguf) | ComfyUI/models/unet | Unet Loader (GGUF) - from the ComfyUI-GGUF node pack |
| Text encoder | ComfyUI/models/text_encoders | Load CLIP / DualCLIPLoader (or the GGUF versions) |
| VAE | ComfyUI/models/vae | Load VAE |
08Measured and reported results
| Label | GPU | Setup | Result | Peak VRAM | Date | Source |
|---|---|---|---|---|---|---|
| measured | RTX 5060 Ti 16 GB | flux-2-klein-9b-Q8_0.gguf · 1024x1024 · 4 steps 3 timed runs after 1 warm-up (19.59 s incl. loading). Text encoder Qwen3 8B FP8. ComfyUI 0.25.0 portable, PyTorch 2.12 cu130, 32 GB RAM. 1024x1024, 4 steps, euler, Flux2Scheduler, | 8.52 s / image | 13.5 GB | 2026-10-01 | |
| measured | RTX 5060 Ti 16 GB | flux-2-klein-9b-Q4_K_M.gguf · 1024x1024 · 4 steps 3 timed runs after 1 warm-up (17.11 s incl. loading). Text encoder Qwen3 8B FP8. ComfyUI 0.25.0 portable, PyTorch 2.12 cu130, 32 GB RAM. 1024x1024, 4 steps, euler, Flux2Scheduler, | 9.85 s / image | 13.5 GB | 2026-10-01 |
09Training a LoRA for FLUX.2 klein 9B
| Trainer | VRAM | Type | Settings and quote | Source |
|---|---|---|---|---|
| OneTrainer | 8 GB | example run | Official preset for klein-base-9B: 512px, batch 2, transformer INT_W8A8, TE fp8, layer offload fraction 0.7, torch.compile on “#flux2 LoRA 8GB.json” | github.com → |
| SimpleTuner | 14 GB | stated minimum | int8 base (~22 GB in bf16); 64GB+ system RAM; single 24GB GPU recommended “klein-9b: ~22GB: ~14GB: 64GB+” | github.com → |
reported From each trainer's own docs, 2026-09-25. Why the figures differ →
10Every file tracked for FLUX.2 klein 9B
| File | Type | Size | Repo |
|---|---|---|---|
| flux-2-klein-9b.safetensors | BF16 | 18.16 GB | black-forest-labs/FLUX.2-klein-9B → |
| flux-2-klein-9b-fp8.safetensors | FP8 | 9.43 GB | black-forest-labs/FLUX.2-klein-9b-fp8 → |
| flux-2-klein-9b-BF16.gguf | BF16 | 18.16 GB | unsloth/FLUX.2-klein-9B-GGUF → |
| flux-2-klein-9b-F16.gguf | F16 | 18.16 GB | unsloth/FLUX.2-klein-9B-GGUF → |
| flux-2-klein-9b-Q8_0.gguf | Q8_0 | 9.98 GB | unsloth/FLUX.2-klein-9B-GGUF → |
| flux-2-klein-9b-Q6_K.gguf | Q6_K | 7.87 GB | unsloth/FLUX.2-klein-9B-GGUF → |
| flux-2-klein-9b-Q5_1.gguf | Q5_1 | 7.25 GB | unsloth/FLUX.2-klein-9B-GGUF → |
| flux-2-klein-9b-Q5_K_M.gguf | Q5_K_M | 7.02 GB | unsloth/FLUX.2-klein-9B-GGUF → |
| flux-2-klein-9b-Q5_K_S.gguf | Q5_K_S | 6.94 GB | unsloth/FLUX.2-klein-9B-GGUF → |
| flux-2-klein-9b-Q5_0.gguf | Q5_0 | 6.71 GB | unsloth/FLUX.2-klein-9B-GGUF → |
| flux-2-klein-9b-Q4_1.gguf | Q4_1 | 6.16 GB | unsloth/FLUX.2-klein-9B-GGUF → |
| flux-2-klein-9b-Q4_K_M.gguf | Q4_K_M | 5.91 GB | unsloth/FLUX.2-klein-9B-GGUF → |
| flux-2-klein-9b-Q4_K_S.gguf | Q4_K_S | 5.83 GB | unsloth/FLUX.2-klein-9B-GGUF → |
| flux-2-klein-9b-Q4_0.gguf | Q4_0 | 5.62 GB | unsloth/FLUX.2-klein-9B-GGUF → |
| flux-2-klein-9b-Q3_K_M.gguf | Q3_K_M | 4.77 GB | unsloth/FLUX.2-klein-9B-GGUF → |
| flux-2-klein-9b-Q3_K_S.gguf | Q3_K_S | 4.69 GB | unsloth/FLUX.2-klein-9B-GGUF → |
| flux-2-klein-9b-Q2_K.gguf | Q2_K | 3.98 GB | unsloth/FLUX.2-klein-9B-GGUF → |
EXTRA (2026 release, very popular). Distilled 4-step, text-to-image + editing. Uses a Qwen3 text encoder (8B class per BFL; size not fetched).