Ming-Image 0.1 Design VRAM requirements
inclusionAI's 6B model from September 2026 for design work: posters, UI mock-ups and text-heavy layouts, with transparent (RGBA) output. The image model is small, but its text encoder is a big language model (12.8–36.7 GB), so system RAM matters as much as VRAM.
01Measured on my RTX 5060 Ti 16 GB
02The files, and how much VRAM each needs
| File | Size | Needed | Min. VRAM | Quality | Source |
|---|---|---|---|---|---|
| 16-bit | 12.3 GB | 14.6 GB | 15 GB | the original weights | Comfy-Org/Ming-Image → |
| Q8_0 | 7.3 GB | 9.6 GB | 10 GB | practically identical to the original | realrebelai/Ming-Image_GGUFs → |
| INT8 | 6.2 GB | 8.5 GB | 9 GB | practically identical to the original | Comfy-Org/Ming-Image → |
| Q6_K | 6.8 GB | 9.1 GB | 10 GB | very close to the original | realrebelai/Ming-Image_GGUFs → |
| Q5_K_M | 6.2 GB | 8.5 GB | 9 GB | close; small differences in fine detail | realrebelai/Ming-Image_GGUFs → |
| Q4_K_M | 5.9 GB | 8.2 GB | 9 GB | good; some loss in fine detail and text | realrebelai/Ming-Image_GGUFs → |
| Q3_K_M | 5.1 GB | 7.4 GB | 8 GB | noticeable loss of detail | realrebelai/Ming-Image_GGUFs → |
| Q2_K | 4.7 GB | 7.0 GB | 8 GB | heavy loss; a last resort | realrebelai/Ming-Image_GGUFs → |
“Needed” = file + 1.5 GB working memory + 0.8 GB system reserve. Made for 2048×2048 as well as 1024×1024; at 2048 the working memory is several times larger than the figure used here. A separate "layer" checkpoint of the same size splits a design into layers. Native in ComfyUI since v0.38.0 (late September 2026).
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 Ming-Image
The cheapest cards (by launch price) that run it well, and every card sorted by memory: best GPU for Ming-Image → 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 | 7.4 GB |
| RTX 3050 6 GB | 6 GB | Offload only | Q3_K_M | 7.4 GB |
| RTX 2070 Super 8 GB | 8 GB | Tight | Q3_K_M | 7.4 GB |
| RTX 2080 Super 8 GB | 8 GB | Tight | Q3_K_M | 7.4 GB |
| RTX 3050 8 GB | 8 GB | Tight | Q3_K_M | 7.4 GB |
| RTX 3060 8 GB | 8 GB | Tight | Q3_K_M | 7.4 GB |
| RTX 3060 Ti 8 GB | 8 GB | Tight | Q3_K_M | 7.4 GB |
| RTX 3070 8 GB | 8 GB | Tight | Q3_K_M | 7.4 GB |
| RTX 3070 Ti 8 GB | 8 GB | Tight | Q3_K_M | 7.4 GB |
| RTX 4060 8 GB | 8 GB | Tight | Q3_K_M | 7.4 GB |
| RTX 4060 Ti 8 GB | 8 GB | Tight | Q3_K_M | 7.4 GB |
| RTX 5050 8 GB | 8 GB | Tight | Q3_K_M | 7.4 GB |
| RTX 5060 8 GB | 8 GB | Tight | Q3_K_M | 7.4 GB |
| RTX 5060 Ti 8 GB | 8 GB | Tight | Q3_K_M | 7.4 GB |
| RX 7600 8 GB | 8 GB | Tight | Q3_K_M | 7.4 GB |
| RX 9050 8 GB | 8 GB | Tight | Q3_K_M | 7.4 GB |
| RX 9060 XT 8 GB | 8 GB | Tight | Q3_K_M | 7.4 GB |
| Arc B570 10 GB | 10 GB | Runs well | Q8_0 | 9.6 GB |
| RTX 3080 10 GB | 10 GB | Runs well | Q8_0 | 9.6 GB |
| RTX 2080 Ti 11 GB | 11 GB | Runs well | Q8_0 | 9.6 GB |
| Arc B580 12 GB | 12 GB | Runs well | Q8_0 | 9.6 GB |
| RTX 2060 12 GB | 12 GB | Runs well | Q8_0 | 9.6 GB |
| RTX 3060 12 GB | 12 GB | Runs well | Q8_0 | 9.6 GB |
| RTX 3080 12 GB | 12 GB | Runs well | Q8_0 | 9.6 GB |
| RTX 3080 Ti 12 GB | 12 GB | Runs well | Q8_0 | 9.6 GB |
| RTX 4070 12 GB | 12 GB | Runs well | INT8 | 8.5 GB |
| RTX 4070 Super 12 GB | 12 GB | Runs well | INT8 | 8.5 GB |
| RTX 4070 Ti 12 GB | 12 GB | Runs well | INT8 | 8.5 GB |
| RTX 5070 12 GB | 12 GB | Runs well | INT8 | 8.5 GB |
| RX 7700 XT 12 GB | 12 GB | Runs well | Q8_0 | 9.6 GB |
| RX 9070 GRE 12 GB | 12 GB | Runs well | INT8 | 8.5 GB |
| Arc A770 16 GB | 16 GB | Runs well | 16-bit | 14.6 GB |
| RTX 4060 Ti 16 GB | 16 GB | Runs well | 16-bit | 14.6 GB |
| RTX 4070 Ti Super 16 GB | 16 GB | Runs well | 16-bit | 14.6 GB |
| RTX 4080 16 GB | 16 GB | Runs well | 16-bit | 14.6 GB |
| RTX 4080 Super 16 GB | 16 GB | Runs well | 16-bit | 14.6 GB |
| RTX 5060 Ti 16 GB | 16 GB | Runs well | 16-bit | 14.6 GB |
| RTX 5070 Ti 16 GB | 16 GB | Runs well | 16-bit | 14.6 GB |
| RTX 5080 16 GB | 16 GB | Runs well | 16-bit | 14.6 GB |
| RX 7600 XT 16 GB | 16 GB | Runs well | 16-bit | 14.6 GB |
| RX 7800 XT 16 GB | 16 GB | Runs well | 16-bit | 14.6 GB |
| RX 7900 GRE 16 GB | 16 GB | Runs well | 16-bit | 14.6 GB |
| RX 9060 XT 16 GB | 16 GB | Runs well | 16-bit | 14.6 GB |
| RX 9070 16 GB | 16 GB | Runs well | 16-bit | 14.6 GB |
| RX 9070 XT 16 GB | 16 GB | Runs well | 16-bit | 14.6 GB |
| RX 7900 XT 20 GB | 20 GB | Runs well | 16-bit | 14.6 GB |
| Arc Pro B60 24 GB | 24 GB | Runs well | 16-bit | 14.6 GB |
| RTX 3090 24 GB | 24 GB | Runs well | 16-bit | 14.6 GB |
| RTX 3090 Ti 24 GB | 24 GB | Runs well | 16-bit | 14.6 GB |
| RTX 4090 24 GB | 24 GB | Runs well | 16-bit | 14.6 GB |
| RX 7900 XTX 24 GB | 24 GB | Runs well | 16-bit | 14.6 GB |
| Arc Pro B70 32 GB | 32 GB | Runs well | 16-bit | 14.6 GB |
| RTX 5090 32 GB | 32 GB | Runs well | 16-bit | 14.6 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 | 7.4 GB |
| RTX 3050 Laptop 6 GB | 6 GB | Offload only | Q3_K_M | 7.4 GB |
| RTX 3060 Laptop 6 GB | 6 GB | Offload only | Q3_K_M | 7.4 GB |
| RTX 4050 Laptop 6 GB | 6 GB | Offload only | Q3_K_M | 7.4 GB |
| RTX 2070 Laptop 8 GB | 8 GB | Tight | Q3_K_M | 7.4 GB |
| RTX 2070 Super Laptop 8 GB | 8 GB | Tight | Q3_K_M | 7.4 GB |
| RTX 2080 Laptop 8 GB | 8 GB | Tight | Q3_K_M | 7.4 GB |
| RTX 2080 Super Laptop 8 GB | 8 GB | Tight | Q3_K_M | 7.4 GB |
| RTX 3070 Laptop 8 GB | 8 GB | Tight | Q3_K_M | 7.4 GB |
| RTX 3070 Ti Laptop 8 GB | 8 GB | Tight | Q3_K_M | 7.4 GB |
| RTX 3080 Laptop 8 GB | 8 GB | Tight | Q3_K_M | 7.4 GB |
| RTX 4060 Laptop 8 GB | 8 GB | Tight | Q3_K_M | 7.4 GB |
| RTX 4070 Laptop 8 GB | 8 GB | Tight | Q3_K_M | 7.4 GB |
| RTX 5050 Laptop 8 GB | 8 GB | Tight | Q3_K_M | 7.4 GB |
| RTX 5060 Laptop 8 GB | 8 GB | Tight | Q3_K_M | 7.4 GB |
| RTX 5070 Laptop 8 GB | 8 GB | Tight | Q3_K_M | 7.4 GB |
| RX 7600M 8 GB | 8 GB | Tight | Q3_K_M | 7.4 GB |
| RX 7600M XT 8 GB | 8 GB | Tight | Q3_K_M | 7.4 GB |
| RX 7600S 8 GB | 8 GB | Tight | Q3_K_M | 7.4 GB |
| RX 7700S 8 GB | 8 GB | Tight | Q3_K_M | 7.4 GB |
| RTX 4080 Laptop 12 GB | 12 GB | Runs well | INT8 | 8.5 GB |
| RTX 5070 Laptop 12 GB | 12 GB | Runs well | INT8 | 8.5 GB |
| RTX 5070 Ti Laptop 12 GB | 12 GB | Runs well | INT8 | 8.5 GB |
| RX 7800M 12 GB | 12 GB | Runs well | Q8_0 | 9.6 GB |
| RTX 3080 Laptop 16 GB | 16 GB | Runs well | 16-bit | 14.6 GB |
| RTX 3080 Ti Laptop 16 GB | 16 GB | Runs well | 16-bit | 14.6 GB |
| RTX 4090 Laptop 16 GB | 16 GB | Runs well | 16-bit | 14.6 GB |
| RTX 5080 Laptop 16 GB | 16 GB | Runs well | 16-bit | 14.6 GB |
| RX 7900M 16 GB | 16 GB | Runs well | 16-bit | 14.6 GB |
| RTX 5090 Laptop 24 GB | 24 GB | Runs well | 16-bit | 14.6 GB |
| Unified memory | ||||
| Radeon 8060S (Strix Halo) 96 GB | 96 GB | Runs well | 16-bit | 14.6 GB |
| Radeon 8065S (Gorgon Halo) 160 GB | 160 GB | Runs well | 16-bit | 14.6 GB |
| Apple Silicon Macs (by memory) | ||||
| Mac 16 GB | 12.7 GB | Runs well | Q8_0 | 9.6 GB |
| Mac 18 GB | 14.4 GB | Runs well | Q8_0 | 9.6 GB |
| Mac 24 GB | 19.6 GB | Runs well | 16-bit | 14.6 GB |
| Mac 32 GB | 26.8 GB | Runs well | 16-bit | 14.6 GB |
| Mac 36 GB | 30.2 GB | Runs well | 16-bit | 14.6 GB |
| Mac 48 GB | 40.2 GB | Runs well | 16-bit | 14.6 GB |
| Mac 64 GB | 55.7 GB | Runs well | 16-bit | 14.6 GB |
| Mac 96 GB | 85 GB | Runs well | 16-bit | 14.6 GB |
| Mac 128 GB | 115.4 GB | Runs well | 16-bit | 14.6 GB |
| Mac 192 GB | 175.4 GB | Runs well | 16-bit | 14.6 GB |
| Mac 256 GB | 236.9 GB | Runs well | 16-bit | 14.6 GB |
| Mac 512 GB | 498.1 GB | Runs well | 16-bit | 14.6 GB |
06Text encoder, VAE and other files
| File | Folder | Size | When | |
|---|---|---|---|---|
| Ling-Mini-2.0 BF16 ming_image_0.1_ling_mini_2.0_bf16.safetensors | models/text_encoders | 36.7 GB | text encoder · alternative | Download → |
| Ling-Mini-2.0 INT8 ming_image_0.1_ling_mini_2.0_int8_convrot.safetensors | models/text_encoders | 19.5 GB | text encoder · smaller, recommended | Download → |
| Ling-Mini-2.0 (4-bit w4a8) ming_image_0.1_ling_mini_2.0_w4a8.safetensors | models/text_encoders | 12.8 GB | text encoder · alternative | Download → |
| Ming-Image VAE ming_image_vae_bf16.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).
Take the smallest Ling-Mini-2.0 (a large language model) file (12.8 GB): it is the only one that fits in 16 GB on its own. Sizes: 16-bit 36.7 GB, INT8 19.5 GB, 4-bit w4a8 12.8 GB (GGUF versions from realrebelai: Q4_K_M 11.5 GB, Q2_K 7.9 GB). Why the encoder does not count against VRAM →
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 | ming_image_0.1_design_int8_convrot.safetensors · 1024x1024 · 12 steps 2 timed runs after 1 warm-up (17.57 s incl. loading). Design variant. Text encoder Ling-Mini-2.0 INT8 (ming_image_0.1_ling_mini_2.0_int8_convrot), Ming-Image VAE. 12 steps, euler/s | 7.42 s / image | 13.6 GB | 2026-10-04 | |
| measured | RTX 5060 Ti 16 GB | ming_image_0.1_design_int8_convrot.safetensors · 2048x2048 · 12 steps 2 timed runs after 1 warm-up (52.11 s incl. loading). Design variant. Text encoder Ling-Mini-2.0 INT8 (ming_image_0.1_ling_mini_2.0_int8_convrot), Ming-Image VAE. 12 steps, euler/s | 44.85 s / image | 14.4 GB | 2026-10-04 | |
| measured | RTX 5060 Ti 16 GB | ming_image_0.1_design_bf16.safetensors · 1024x1024 · 12 steps 2 timed runs after 1 warm-up (28.3 s incl. loading). Design variant. Text encoder Ling-Mini-2.0 INT8 (ming_image_0.1_ling_mini_2.0_int8_convrot), Ming-Image VAE. 12 steps, euler/si | 17.39 s / image | 15.3 GB | 2026-10-04 |
09Training a LoRA for Ming-Image
No trainer documentation with a VRAM figure for this model was found yet. What the trainers say for other models →
10Every file tracked for Ming-Image
| File | Type | Size | Repo |
|---|---|---|---|
| ming_image_0.1_design_layer_bf16.safetensors (layer) | BF16 | 12.31 GB | Comfy-Org/Ming-Image → |
| ming_image_0.1_design_bf16.safetensors (design) | BF16 | 12.31 GB | Comfy-Org/Ming-Image → |
| ming_image_0.1_design_layer_int8_convrot.safetensors (layer) | INT8 | 6.18 GB | Comfy-Org/Ming-Image → |
| ming_image_0.1_design_int8_convrot.safetensors (design) | INT8 | 6.18 GB | Comfy-Org/Ming-Image → |
| ming-image-0.1-design-layer-nvfp4.gguf (layer) | NVFP4 | 3.48 GB | gguf-org/ming-image-gguf → |
| ming-image-0.1-design-nvfp4.gguf (design) | NVFP4 | 3.48 GB | gguf-org/ming-image-gguf → |
| Ming-Image-0.1-Design-Q8_0.gguf (design) | Q8_0 | 7.33 GB | realrebelai/Ming-Image_GGUFs → |
| Ming-Image-0.1-Design-Q6_K.gguf (design) | Q6_K | 6.76 GB | realrebelai/Ming-Image_GGUFs → |
| Ming-Image-0.1-Design-Q5_K_M.gguf (design) | Q5_K_M | 6.16 GB | realrebelai/Ming-Image_GGUFs → |
| Ming-Image-0.1-Design-Q4_K_M.gguf (design) | Q4_K_M | 5.87 GB | realrebelai/Ming-Image_GGUFs → |
| Ming-Image-0.1-Design-Q3_K_M.gguf (design) | Q3_K_M | 5.13 GB | realrebelai/Ming-Image_GGUFs → |
| Ming-Image-0.1-Design-Q2_K.gguf (design) | Q2_K | 4.71 GB | realrebelai/Ming-Image_GGUFs → |
6B DiT (30 layers) for text-heavy design: UI, posters, infographics; native RGBA/transparent output; 1024 and 2048 native res (2048 recommended); 12 steps, CFG 1.0. A separate 'layer' checkpoint (same size) does layer decomposition. The text encoder is very large (Ling-Mini-2.0 MoE-based, 36.7 GB in bf16), which dominates memory needs; the official repo validated on one 80 GiB GPU. Comfy-Org also hosts a "layer" checkpoint (same size) for splitting a design into layers.