Qwen-Image 2.1 VRAM requirements
September 2026: a new 7B model that does text-to-image and editing in one, with transparent-background (RGBA) output. Far lighter than the 20B Qwen-Image.
TypeImage
Parameters7.1B
Fits entirely from6 GB
8-bit or better from10 GB
01Measured on my RTX 5060 Ti 16 GB
INT8 · 1024×1024FASTEST
17.2s / image
qwen_image_2.1_int8_convrot.safetensors
Peak VRAM14.3 GB
Run range17.1–17.2 s
16-bit · 1024×1024
39.7s / image
qwen_image_2.1_bf16.safetensors
Peak VRAM15.5 GB
Run range39.7–39.7 s
Q8_0 · 1024×1024
43.6s / image
qwen_image_2.1_Q8_0.gguf
Peak VRAM14.8 GB
Run range43.5–43.6 s
02The files, and how much VRAM each needs
| File | Size | Needed | Min. VRAM | Quality | Source |
|---|---|---|---|---|---|
| 16-bit | 14.2 GB | 16.5 GB | 17 GB | the original weights | Comfy-Org/Qwen-Image-2.1 → |
| Q8_0 | 7.6 GB | 9.9 GB | 10 GB | practically identical to the original | Abiray/Qwen-Image-2.1-GGUF → |
| INT8 | 7.3 GB | 9.6 GB | 10 GB | practically identical to the original | Comfy-Org/Qwen-Image-2.1 → |
| Q6_K | 5.9 GB | 8.2 GB | 9 GB | very close to the original | Abiray/Qwen-Image-2.1-GGUF → |
| Q5_K_M | 5.0 GB | 7.3 GB | 8 GB | close; small differences in fine detail | Abiray/Qwen-Image-2.1-GGUF → |
| Q4_K_M | 4.2 GB | 6.5 GB | 7 GB | good; some loss in fine detail and text | Abiray/Qwen-Image-2.1-GGUF → |
| Q3_K_M | 3.2 GB | 5.5 GB | 6 GB | noticeable loss of detail | Abiray/Qwen-Image-2.1-GGUF → |
“Needed” = file + 1.5 GB working memory + 0.8 GB system reserve.
Free ComfyUI workflow, tested on my RTX 5060 Ti: INT8 workflow (.json). Drag the file into ComfyUI. All workflows →
03By amount of VRAM
04Best GPU for Qwen-Image 2.1
The cheapest cards (by launch price) that run it well, and every card sorted by memory: best GPU for Qwen-Image 2.1 → 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 | Tight | Q3_K_M | 5.5 GB |
| RTX 3050 6 GB | 6 GB | Tight | Q3_K_M | 5.5 GB |
| RTX 2070 Super 8 GB | 8 GB | Runs | Q5_K_M | 7.3 GB |
| RTX 2080 Super 8 GB | 8 GB | Runs | Q5_K_M | 7.3 GB |
| RTX 3050 8 GB | 8 GB | Runs | Q5_K_M | 7.3 GB |
| RTX 3060 8 GB | 8 GB | Runs | Q5_K_M | 7.3 GB |
| RTX 3060 Ti 8 GB | 8 GB | Runs | Q5_K_M | 7.3 GB |
| RTX 3070 8 GB | 8 GB | Runs | Q5_K_M | 7.3 GB |
| RTX 3070 Ti 8 GB | 8 GB | Runs | Q5_K_M | 7.3 GB |
| RTX 4060 8 GB | 8 GB | Runs | Q5_K_M | 7.3 GB |
| RTX 4060 Ti 8 GB | 8 GB | Runs | Q5_K_M | 7.3 GB |
| RTX 5050 8 GB | 8 GB | Runs | Q5_K_M | 7.3 GB |
| RTX 5060 8 GB | 8 GB | Runs | Q5_K_M | 7.3 GB |
| RTX 5060 Ti 8 GB | 8 GB | Runs | Q5_K_M | 7.3 GB |
| RX 7600 8 GB | 8 GB | Runs | Q5_K_M | 7.3 GB |
| RX 9050 8 GB | 8 GB | Runs | Q5_K_M | 7.3 GB |
| RX 9060 XT 8 GB | 8 GB | Runs | Q5_K_M | 7.3 GB |
| Arc B570 10 GB | 10 GB | Runs well | Q8_0 | 9.9 GB |
| RTX 3080 10 GB | 10 GB | Runs well | Q8_0 | 9.9 GB |
| RTX 2080 Ti 11 GB | 11 GB | Runs well | Q8_0 | 9.9 GB |
| Arc B580 12 GB | 12 GB | Runs well | Q8_0 | 9.9 GB |
| RTX 2060 12 GB | 12 GB | Runs well | Q8_0 | 9.9 GB |
| RTX 3060 12 GB | 12 GB | Runs well | Q8_0 | 9.9 GB |
| RTX 3080 12 GB | 12 GB | Runs well | Q8_0 | 9.9 GB |
| RTX 3080 Ti 12 GB | 12 GB | Runs well | Q8_0 | 9.9 GB |
| RTX 4070 12 GB | 12 GB | Runs well | INT8 | 9.6 GB |
| RTX 4070 Super 12 GB | 12 GB | Runs well | INT8 | 9.6 GB |
| RTX 4070 Ti 12 GB | 12 GB | Runs well | INT8 | 9.6 GB |
| RTX 5070 12 GB | 12 GB | Runs well | INT8 | 9.6 GB |
| RX 7700 XT 12 GB | 12 GB | Runs well | Q8_0 | 9.9 GB |
| RX 9070 GRE 12 GB | 12 GB | Runs well | INT8 | 9.6 GB |
| Arc A770 16 GB | 16 GB | Runs well | Q8_0 | 9.9 GB |
| RTX 4060 Ti 16 GB | 16 GB | Runs well | INT8 | 9.6 GB |
| RTX 4070 Ti Super 16 GB | 16 GB | Runs well | INT8 | 9.6 GB |
| RTX 4080 16 GB | 16 GB | Runs well | INT8 | 9.6 GB |
| RTX 4080 Super 16 GB | 16 GB | Runs well | INT8 | 9.6 GB |
| RTX 5060 Ti 16 GB | 16 GB | Runs well | INT8 | 9.6 GB |
| RTX 5070 Ti 16 GB | 16 GB | Runs well | INT8 | 9.6 GB |
| RTX 5080 16 GB | 16 GB | Runs well | INT8 | 9.6 GB |
| RX 7600 XT 16 GB | 16 GB | Runs well | Q8_0 | 9.9 GB |
| RX 7800 XT 16 GB | 16 GB | Runs well | Q8_0 | 9.9 GB |
| RX 7900 GRE 16 GB | 16 GB | Runs well | Q8_0 | 9.9 GB |
| RX 9060 XT 16 GB | 16 GB | Runs well | INT8 | 9.6 GB |
| RX 9070 16 GB | 16 GB | Runs well | INT8 | 9.6 GB |
| RX 9070 XT 16 GB | 16 GB | Runs well | INT8 | 9.6 GB |
| RX 7900 XT 20 GB | 20 GB | Runs well | 16-bit | 16.5 GB |
| Arc Pro B60 24 GB | 24 GB | Runs well | 16-bit | 16.5 GB |
| RTX 3090 24 GB | 24 GB | Runs well | 16-bit | 16.5 GB |
| RTX 3090 Ti 24 GB | 24 GB | Runs well | 16-bit | 16.5 GB |
| RTX 4090 24 GB | 24 GB | Runs well | 16-bit | 16.5 GB |
| RX 7900 XTX 24 GB | 24 GB | Runs well | 16-bit | 16.5 GB |
| Arc Pro B70 32 GB | 32 GB | Runs well | 16-bit | 16.5 GB |
| RTX 5090 32 GB | 32 GB | Runs well | 16-bit | 16.5 GB |
| Laptop GPUs | ||||
| RTX 3050 Laptop 4 GB | 4 GB | Offload only | Q3_K_M | 5.5 GB |
| RTX 3050 Ti Laptop 4 GB | 4 GB | Offload only | Q3_K_M | 5.5 GB |
| RTX 2060 Laptop 6 GB | 6 GB | Tight | Q3_K_M | 5.5 GB |
| RTX 3050 Laptop 6 GB | 6 GB | Tight | Q3_K_M | 5.5 GB |
| RTX 3060 Laptop 6 GB | 6 GB | Tight | Q3_K_M | 5.5 GB |
| RTX 4050 Laptop 6 GB | 6 GB | Tight | Q3_K_M | 5.5 GB |
| RTX 2070 Laptop 8 GB | 8 GB | Runs | Q5_K_M | 7.3 GB |
| RTX 2070 Super Laptop 8 GB | 8 GB | Runs | Q5_K_M | 7.3 GB |
| RTX 2080 Laptop 8 GB | 8 GB | Runs | Q5_K_M | 7.3 GB |
| RTX 2080 Super Laptop 8 GB | 8 GB | Runs | Q5_K_M | 7.3 GB |
| RTX 3070 Laptop 8 GB | 8 GB | Runs | Q5_K_M | 7.3 GB |
| RTX 3070 Ti Laptop 8 GB | 8 GB | Runs | Q5_K_M | 7.3 GB |
| RTX 3080 Laptop 8 GB | 8 GB | Runs | Q5_K_M | 7.3 GB |
| RTX 4060 Laptop 8 GB | 8 GB | Runs | Q5_K_M | 7.3 GB |
| RTX 4070 Laptop 8 GB | 8 GB | Runs | Q5_K_M | 7.3 GB |
| RTX 5050 Laptop 8 GB | 8 GB | Runs | Q5_K_M | 7.3 GB |
| RTX 5060 Laptop 8 GB | 8 GB | Runs | Q5_K_M | 7.3 GB |
| RTX 5070 Laptop 8 GB | 8 GB | Runs | Q5_K_M | 7.3 GB |
| RX 7600M 8 GB | 8 GB | Runs | Q5_K_M | 7.3 GB |
| RX 7600M XT 8 GB | 8 GB | Runs | Q5_K_M | 7.3 GB |
| RX 7600S 8 GB | 8 GB | Runs | Q5_K_M | 7.3 GB |
| RX 7700S 8 GB | 8 GB | Runs | Q5_K_M | 7.3 GB |
| RTX 4080 Laptop 12 GB | 12 GB | Runs well | INT8 | 9.6 GB |
| RTX 5070 Laptop 12 GB | 12 GB | Runs well | INT8 | 9.6 GB |
| RTX 5070 Ti Laptop 12 GB | 12 GB | Runs well | INT8 | 9.6 GB |
| RX 7800M 12 GB | 12 GB | Runs well | Q8_0 | 9.9 GB |
| RTX 3080 Laptop 16 GB | 16 GB | Runs well | Q8_0 | 9.9 GB |
| RTX 3080 Ti Laptop 16 GB | 16 GB | Runs well | Q8_0 | 9.9 GB |
| RTX 4090 Laptop 16 GB | 16 GB | Runs well | INT8 | 9.6 GB |
| RTX 5080 Laptop 16 GB | 16 GB | Runs well | INT8 | 9.6 GB |
| RX 7900M 16 GB | 16 GB | Runs well | Q8_0 | 9.9 GB |
| RTX 5090 Laptop 24 GB | 24 GB | Runs well | 16-bit | 16.5 GB |
| Unified memory | ||||
| Radeon 8060S (Strix Halo) 96 GB | 96 GB | Runs well | 16-bit | 16.5 GB |
| Radeon 8065S (Gorgon Halo) 160 GB | 160 GB | Runs well | 16-bit | 16.5 GB |
| Apple Silicon Macs (by memory) | ||||
| Mac 16 GB | 12.7 GB | Runs well | Q8_0 | 9.9 GB |
| Mac 18 GB | 14.4 GB | Runs well | Q8_0 | 9.9 GB |
| Mac 24 GB | 19.6 GB | Runs well | 16-bit | 16.5 GB |
| Mac 32 GB | 26.8 GB | Runs well | 16-bit | 16.5 GB |
| Mac 36 GB | 30.2 GB | Runs well | 16-bit | 16.5 GB |
| Mac 48 GB | 40.2 GB | Runs well | 16-bit | 16.5 GB |
| Mac 64 GB | 55.7 GB | Runs well | 16-bit | 16.5 GB |
| Mac 96 GB | 85 GB | Runs well | 16-bit | 16.5 GB |
| Mac 128 GB | 115.4 GB | Runs well | 16-bit | 16.5 GB |
| Mac 192 GB | 175.4 GB | Runs well | 16-bit | 16.5 GB |
| Mac 256 GB | 236.9 GB | Runs well | 16-bit | 16.5 GB |
| Mac 512 GB | 498.1 GB | Runs well | 16-bit | 16.5 GB |
06Text encoder, VAE and other files
| File | Folder | Size | When | |
|---|---|---|---|---|
| Qwen3-VL 8B INT8 qwen3vl_8b_int8_convrot.safetensors | models/text_encoders | 9.4 GB | text encoder · smaller, recommended | Download → |
| Qwen3-VL 8B BF16 qwen3vl_8b_bf16.safetensors | models/text_encoders | 17.5 GB | text encoder · alternative | Download → |
| Qwen-Image 2.1 VAE qwen_image_2.1_vae_bf16.safetensors | models/vae | 0.7 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 | qwen_image_2.1_int8_convrot.safetensors · 1024x1024 · 25 steps 2 timed runs after 1 warm-up (24.6 s incl. loading). Text encoder Qwen3-VL 8B INT8 (qwen3vl_8b_int8_convrot), Qwen-Image 2.1 VAE. 25 steps, euler/simple, CFG 1, TextEncodeQwenImage | 17.16 s / image | 14.3 GB | 2026-10-04 | |
| measured | RTX 5060 Ti 16 GB | qwen_image_2.1_Q8_0.gguf · 1024x1024 · 25 steps 2 timed runs after 1 warm-up (53.91 s incl. loading). Text encoder Qwen3-VL 8B INT8 (qwen3vl_8b_int8_convrot), Qwen-Image 2.1 VAE. 25 steps, euler/simple, CFG 1, TextEncodeQwenImag | 43.57 s / image | 14.8 GB | 2026-10-04 | |
| measured | RTX 5060 Ti 16 GB | qwen_image_2.1_bf16.safetensors · 1024x1024 · 25 steps 2 timed runs after 1 warm-up (46.2 s incl. loading). Text encoder Qwen3-VL 8B INT8 (qwen3vl_8b_int8_convrot), Qwen-Image 2.1 VAE. 25 steps, euler/simple, CFG 1, TextEncodeQwenImage | 39.67 s / image | 15.5 GB | 2026-10-04 | |
| reported | RTX 4060 Ti 16 GB | Qwen Image 2.1 INT8 Convrot (+Qwen3 VL 8B INT8 TE) “Image generation: ~20 seconds ... Peak VRAM: ~15 GB” ComfyUI, no CPU/disk offload; approximate values ('~'); editing ~60 s; resolution/steps not stated; date derived from '1 day ago' on 2026-09-25 | 20 s / image | 15 GB | 2026-09-24 | huggingface.co → |
| reported | RTX 4070 12 GB | Qwen-Image-2.1 int8 convrot + qwen3vl_8b_int8 convrot TE + VAE bf16 · 1024x1024 · 25 steps “1024x1024・25ステップ（公式ワークフロー相当） | 14.0 秒” ComfyUI, RTX 4070 12GB with partial offload (17GB weights); same table: 12 steps 9.6 s, 2048x2048 20 steps 90.5 s; repo date not shown (Qwen-Image 2.1 era, 2026) | 14 s / image | — | — | github.com → |
| reported | RTX 4070 12 GB | Qwen Image 2.1 INT8 ConvRot · 832x1248 · 25 steps “「全体の所要時間」で見ると、平均で32.63秒から19.83秒へ（約39.2%短縮）… GPU全体のVRAMピーク：11.19 GiB → 11.28 GiB” ComfyUI, RTX 4070 12GB, Euler/simple CFG 1; 32.63 s = standard total time (19.83 s with Spectrum speedup node); VRAM in GiB | 32.63 s / image | 11.19 GB | 2026-09-24 | note.com → |
Other people's numbers are copied as published, with a link. Send yours.
09Training a LoRA for Qwen-Image 2.1
No trainer documentation with a VRAM figure for this model was found yet. What the trainers say for other models →
10Every file tracked for Qwen-Image 2.1
| File | Type | Size | Repo |
|---|---|---|---|
| qwen_image_2.1_Q8_0.gguf | Q8_0 | 7.59 GB | Abiray/Qwen-Image-2.1-GGUF → |
| qwen_image_2.1_Q6_K.gguf | Q6_K | 5.88 GB | Abiray/Qwen-Image-2.1-GGUF → |
| qwen_image_2.1_Q5_K_M.gguf | Q5_K_M | 5.01 GB | Abiray/Qwen-Image-2.1-GGUF → |
| qwen_image_2.1_Q4_K_M.gguf | Q4_K_M | 4.19 GB | Abiray/Qwen-Image-2.1-GGUF → |
| qwen_image_2.1_Q4_K_S.gguf | Q4_K_S | 4.06 GB | Abiray/Qwen-Image-2.1-GGUF → |
| qwen_image_2.1_Q3_K_M.gguf | Q3_K_M | 3.19 GB | Abiray/Qwen-Image-2.1-GGUF → |
| qwen_image_2.1_bf16.safetensors | BF16 | 14.23 GB | Comfy-Org/Qwen-Image-2.1 → |
| qwen_image_2.1_int8_convrot.safetensors | INT8 | 7.26 GB | Comfy-Org/Qwen-Image-2.1 → |
| qwen-image-2.1-F16.gguf | F16 | 14.23 GB | unsloth/Qwen-Image-2.1-GGUF → |
| qwen-image-2.1-Q8_0.gguf | Q8_0 | 7.64 GB | unsloth/Qwen-Image-2.1-GGUF → |
| qwen-image-2.1-Q6_K_XL.gguf | Q6_K_XL | 6.72 GB | unsloth/Qwen-Image-2.1-GGUF → |
| qwen-image-2.1-Q6_K.gguf | Q6_K | 6.27 GB | unsloth/Qwen-Image-2.1-GGUF → |
| qwen-image-2.1-Q5_K_M.gguf | Q5_K_M | 5.39 GB | unsloth/Qwen-Image-2.1-GGUF → |
| qwen-image-2.1-Q5_K_S.gguf | Q5_K_S | 4.50 GB | unsloth/Qwen-Image-2.1-GGUF → |
| qwen-image-2.1-Q4_K_M.gguf | Q4_K_M | 4.20 GB | unsloth/Qwen-Image-2.1-GGUF → |
| qwen-image-2.1-Q4_K_S.gguf | Q4_K_S | 3.91 GB | unsloth/Qwen-Image-2.1-GGUF → |
| qwen-image-2.1-Q3_K_XL.gguf | Q3_K_XL | 3.61 GB | unsloth/Qwen-Image-2.1-GGUF → |
| qwen-image-2.1-Q3_K_M.gguf | Q3_K_M | 3.17 GB | unsloth/Qwen-Image-2.1-GGUF → |
| qwen-image-2.1-Q3_K_S.gguf | Q3_K_S | 2.72 GB | unsloth/Qwen-Image-2.1-GGUF → |
| qwen-image-2.1-Q2_K.gguf | Q2_K | 2.47 GB | unsloth/Qwen-Image-2.1-GGUF → |
New 7B single-stream DiT (HF API 7.1B) - much smaller than the 20B Qwen-Image; one model does text-to-image AND editing, with native RGBA transparency. Optional prompt-enhancer LLM (Qwen3.5-9B int8_convrot, 9471072252 each for t2i/i2i) used via TextGenerate node. unsloth/Qwen-Image-2.1-FP8 (Qwen-Image-2.1-FP8.safetensors 7122877560) exists but is not documented for ComfyUI, so not listed as a file.