ERNIE-Image (Turbo) VRAM requirements
Baidu's 8B model from April 2026, strong at rendering text in images. Turbo runs in about 8 steps; Base takes 50.
TypeImage
Parameters8.03B
Fits entirely from6 GB
8-bit or better from11 GB
01The files, and how much VRAM each needs
| File | Size | Needed | Min. VRAM | Quality | Source |
|---|---|---|---|---|---|
| 16-bit | 16.1 GB | 18.4 GB | 19 GB | the original weights | Comfy-Org/ERNIE-Image → |
| Q8_0 | 8.7 GB | 11.0 GB | 11 GB | practically identical to the original | unsloth/ERNIE-Image-Turbo-GGUF → |
| Q6_K | 6.8 GB | 9.1 GB | 10 GB | very close to the original | unsloth/ERNIE-Image-Turbo-GGUF → |
| Q5_K_M | 5.9 GB | 8.2 GB | 9 GB | close; small differences in fine detail | unsloth/ERNIE-Image-Turbo-GGUF → |
| Q4_K_M | 5.0 GB | 7.3 GB | 8 GB | good; some loss in fine detail and text | unsloth/ERNIE-Image-Turbo-GGUF → |
| Q3_K_M | 3.9 GB | 6.2 GB | 7 GB | noticeable loss of detail | unsloth/ERNIE-Image-Turbo-GGUF → |
| Q2_K | 3.2 GB | 5.5 GB | 6 GB | heavy loss; a last resort | unsloth/ERNIE-Image-Turbo-GGUF → |
“Needed” = file + 1.5 GB working memory + 0.8 GB system reserve.
02By amount of VRAM
03Best GPU for ERNIE-Image
The cheapest cards (by launch price) that run it well, and every card sorted by memory: best GPU for ERNIE-Image → Planning bigger images or longer clips? Open the calculator →
04By graphics card
| GPU | VRAM | Verdict | Best file | Needed |
|---|---|---|---|---|
| Desktop graphics cards | ||||
| RTX 2060 6 GB | 6 GB | Tight | Q2_K | 5.5 GB |
| RTX 3050 6 GB | 6 GB | Tight | Q2_K | 5.5 GB |
| RTX 2070 Super 8 GB | 8 GB | Runs | Q4_K_M | 7.3 GB |
| RTX 2080 Super 8 GB | 8 GB | Runs | Q4_K_M | 7.3 GB |
| RTX 3050 8 GB | 8 GB | Runs | Q4_K_M | 7.3 GB |
| RTX 3060 8 GB | 8 GB | Runs | Q4_K_M | 7.3 GB |
| RTX 3060 Ti 8 GB | 8 GB | Runs | Q4_K_M | 7.3 GB |
| RTX 3070 8 GB | 8 GB | Runs | Q4_K_M | 7.3 GB |
| RTX 3070 Ti 8 GB | 8 GB | Runs | Q4_K_M | 7.3 GB |
| RTX 4060 8 GB | 8 GB | Runs | Q4_K_M | 7.3 GB |
| RTX 4060 Ti 8 GB | 8 GB | Runs | Q4_K_M | 7.3 GB |
| RTX 5050 8 GB | 8 GB | Runs | Q4_K_M | 7.3 GB |
| RTX 5060 8 GB | 8 GB | Runs | Q4_K_M | 7.3 GB |
| RTX 5060 Ti 8 GB | 8 GB | Runs | Q4_K_M | 7.3 GB |
| RX 7600 8 GB | 8 GB | Runs | Q4_K_M | 7.3 GB |
| RX 9050 8 GB | 8 GB | Runs | Q4_K_M | 7.3 GB |
| RX 9060 XT 8 GB | 8 GB | Runs | Q4_K_M | 7.3 GB |
| Arc B570 10 GB | 10 GB | Runs | Q6_K | 9.1 GB |
| RTX 3080 10 GB | 10 GB | Runs | Q6_K | 9.1 GB |
| RTX 2080 Ti 11 GB | 11 GB | Runs well | Q8_0 | 11.0 GB |
| Arc B580 12 GB | 12 GB | Runs well | Q8_0 | 11.0 GB |
| RTX 2060 12 GB | 12 GB | Runs well | Q8_0 | 11.0 GB |
| RTX 3060 12 GB | 12 GB | Runs well | Q8_0 | 11.0 GB |
| RTX 3080 12 GB | 12 GB | Runs well | Q8_0 | 11.0 GB |
| RTX 3080 Ti 12 GB | 12 GB | Runs well | Q8_0 | 11.0 GB |
| RTX 4070 12 GB | 12 GB | Runs well | Q8_0 | 11.0 GB |
| RTX 4070 Super 12 GB | 12 GB | Runs well | Q8_0 | 11.0 GB |
| RTX 4070 Ti 12 GB | 12 GB | Runs well | Q8_0 | 11.0 GB |
| RTX 5070 12 GB | 12 GB | Runs well | Q8_0 | 11.0 GB |
| RX 7700 XT 12 GB | 12 GB | Runs well | Q8_0 | 11.0 GB |
| RX 9070 GRE 12 GB | 12 GB | Runs well | Q8_0 | 11.0 GB |
| Arc A770 16 GB | 16 GB | Runs well | Q8_0 | 11.0 GB |
| RTX 4060 Ti 16 GB | 16 GB | Runs well | Q8_0 | 11.0 GB |
| RTX 4070 Ti Super 16 GB | 16 GB | Runs well | Q8_0 | 11.0 GB |
| RTX 4080 16 GB | 16 GB | Runs well | Q8_0 | 11.0 GB |
| RTX 4080 Super 16 GB | 16 GB | Runs well | Q8_0 | 11.0 GB |
| RTX 5060 Ti 16 GB | 16 GB | Runs well | Q8_0 | 11.0 GB |
| RTX 5070 Ti 16 GB | 16 GB | Runs well | Q8_0 | 11.0 GB |
| RTX 5080 16 GB | 16 GB | Runs well | Q8_0 | 11.0 GB |
| RX 7600 XT 16 GB | 16 GB | Runs well | Q8_0 | 11.0 GB |
| RX 7800 XT 16 GB | 16 GB | Runs well | Q8_0 | 11.0 GB |
| RX 7900 GRE 16 GB | 16 GB | Runs well | Q8_0 | 11.0 GB |
| RX 9060 XT 16 GB | 16 GB | Runs well | Q8_0 | 11.0 GB |
| RX 9070 16 GB | 16 GB | Runs well | Q8_0 | 11.0 GB |
| RX 9070 XT 16 GB | 16 GB | Runs well | Q8_0 | 11.0 GB |
| RX 7900 XT 20 GB | 20 GB | Runs well | 16-bit | 18.4 GB |
| Arc Pro B60 24 GB | 24 GB | Runs well | 16-bit | 18.4 GB |
| RTX 3090 24 GB | 24 GB | Runs well | 16-bit | 18.4 GB |
| RTX 3090 Ti 24 GB | 24 GB | Runs well | 16-bit | 18.4 GB |
| RTX 4090 24 GB | 24 GB | Runs well | 16-bit | 18.4 GB |
| RX 7900 XTX 24 GB | 24 GB | Runs well | 16-bit | 18.4 GB |
| Arc Pro B70 32 GB | 32 GB | Runs well | 16-bit | 18.4 GB |
| RTX 5090 32 GB | 32 GB | Runs well | 16-bit | 18.4 GB |
| Laptop GPUs | ||||
| RTX 3050 Laptop 4 GB | 4 GB | Offload only | Q2_K | 5.5 GB |
| RTX 3050 Ti Laptop 4 GB | 4 GB | Offload only | Q2_K | 5.5 GB |
| RTX 2060 Laptop 6 GB | 6 GB | Tight | Q2_K | 5.5 GB |
| RTX 3050 Laptop 6 GB | 6 GB | Tight | Q2_K | 5.5 GB |
| RTX 3060 Laptop 6 GB | 6 GB | Tight | Q2_K | 5.5 GB |
| RTX 4050 Laptop 6 GB | 6 GB | Tight | Q2_K | 5.5 GB |
| RTX 2070 Laptop 8 GB | 8 GB | Runs | Q4_K_M | 7.3 GB |
| RTX 2070 Super Laptop 8 GB | 8 GB | Runs | Q4_K_M | 7.3 GB |
| RTX 2080 Laptop 8 GB | 8 GB | Runs | Q4_K_M | 7.3 GB |
| RTX 2080 Super Laptop 8 GB | 8 GB | Runs | Q4_K_M | 7.3 GB |
| RTX 3070 Laptop 8 GB | 8 GB | Runs | Q4_K_M | 7.3 GB |
| RTX 3070 Ti Laptop 8 GB | 8 GB | Runs | Q4_K_M | 7.3 GB |
| RTX 3080 Laptop 8 GB | 8 GB | Runs | Q4_K_M | 7.3 GB |
| RTX 4060 Laptop 8 GB | 8 GB | Runs | Q4_K_M | 7.3 GB |
| RTX 4070 Laptop 8 GB | 8 GB | Runs | Q4_K_M | 7.3 GB |
| RTX 5050 Laptop 8 GB | 8 GB | Runs | Q4_K_M | 7.3 GB |
| RTX 5060 Laptop 8 GB | 8 GB | Runs | Q4_K_M | 7.3 GB |
| RTX 5070 Laptop 8 GB | 8 GB | Runs | Q4_K_M | 7.3 GB |
| RX 7600M 8 GB | 8 GB | Runs | Q4_K_M | 7.3 GB |
| RX 7600M XT 8 GB | 8 GB | Runs | Q4_K_M | 7.3 GB |
| RX 7600S 8 GB | 8 GB | Runs | Q4_K_M | 7.3 GB |
| RX 7700S 8 GB | 8 GB | Runs | Q4_K_M | 7.3 GB |
| RTX 4080 Laptop 12 GB | 12 GB | Runs well | Q8_0 | 11.0 GB |
| RTX 5070 Laptop 12 GB | 12 GB | Runs well | Q8_0 | 11.0 GB |
| RTX 5070 Ti Laptop 12 GB | 12 GB | Runs well | Q8_0 | 11.0 GB |
| RX 7800M 12 GB | 12 GB | Runs well | Q8_0 | 11.0 GB |
| RTX 3080 Laptop 16 GB | 16 GB | Runs well | Q8_0 | 11.0 GB |
| RTX 3080 Ti Laptop 16 GB | 16 GB | Runs well | Q8_0 | 11.0 GB |
| RTX 4090 Laptop 16 GB | 16 GB | Runs well | Q8_0 | 11.0 GB |
| RTX 5080 Laptop 16 GB | 16 GB | Runs well | Q8_0 | 11.0 GB |
| RX 7900M 16 GB | 16 GB | Runs well | Q8_0 | 11.0 GB |
| RTX 5090 Laptop 24 GB | 24 GB | Runs well | 16-bit | 18.4 GB |
| Unified memory | ||||
| Radeon 8060S (Strix Halo) 96 GB | 96 GB | Runs well | 16-bit | 18.4 GB |
| Radeon 8065S (Gorgon Halo) 160 GB | 160 GB | Runs well | 16-bit | 18.4 GB |
| Apple Silicon Macs (by memory) | ||||
| Mac 16 GB | 12.7 GB | Runs well | Q8_0 | 11.0 GB |
| Mac 18 GB | 14.4 GB | Runs well | Q8_0 | 11.0 GB |
| Mac 24 GB | 19.6 GB | Runs well | 16-bit | 18.4 GB |
| Mac 32 GB | 26.8 GB | Runs well | 16-bit | 18.4 GB |
| Mac 36 GB | 30.2 GB | Runs well | 16-bit | 18.4 GB |
| Mac 48 GB | 40.2 GB | Runs well | 16-bit | 18.4 GB |
| Mac 64 GB | 55.7 GB | Runs well | 16-bit | 18.4 GB |
| Mac 96 GB | 85 GB | Runs well | 16-bit | 18.4 GB |
| Mac 128 GB | 115.4 GB | Runs well | 16-bit | 18.4 GB |
| Mac 192 GB | 175.4 GB | Runs well | 16-bit | 18.4 GB |
| Mac 256 GB | 236.9 GB | Runs well | 16-bit | 18.4 GB |
| Mac 512 GB | 498.1 GB | Runs well | 16-bit | 18.4 GB |
05Text encoder, VAE and other files
| File | Folder | Size | When | |
|---|---|---|---|---|
| Ministral 3 3B ministral-3-3b.safetensors | models/text_encoders | 7.7 GB | required | Download → |
| FLUX.2 VAE flux2-vae.safetensors | models/vae | 0.3 GB | required | Download → |
| ERNIE prompt enhancer (optional) ernie-image-prompt-enhancer.safetensors | models/text_encoders | 6.9 GB | optional | Download → |
The files the official ComfyUI workflows load next to the model. Sizes read from Hugging Face (2026-10-01).
06Where 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 |
07Measured and reported results
No measured results yet. Send yours.
08Training a LoRA for ERNIE-Image
No trainer documentation with a VRAM figure for this model was found yet. What the trainers say for other models →
09Every file tracked for ERNIE-Image
| File | Type | Size | Repo |
|---|---|---|---|
| ernie-image-turbo-int8rowwise.safetensors (turbo) | INT8 | 8.22 GB | Bedovyy/ERNIE-Image-Quantized → needs ComfyUI-INT8-Fast custom node |
| ernie-image-int8rowwise.safetensors (base) | INT8 | 8.22 GB | Bedovyy/ERNIE-Image-Quantized → needs ComfyUI-INT8-Fast custom node |
| ernie-image-turbo-fp8.safetensors (turbo) | FP8 | 8.22 GB | Bedovyy/ERNIE-Image-Quantized → fp8 e4m3 |
| ernie-image-fp8.safetensors (base) | FP8 | 8.22 GB | Bedovyy/ERNIE-Image-Quantized → fp8 e4m3 |
| ernie-image-turbo-nvfp4.safetensors (turbo) | NVFP4 | 4.78 GB | Bedovyy/ERNIE-Image-Quantized → |
| ernie-image-nvfp4.safetensors (base) | NVFP4 | 4.78 GB | Bedovyy/ERNIE-Image-Quantized → |
| ernie-image-turbo.safetensors (turbo) | 16-bit | 16.07 GB | Comfy-Org/ERNIE-Image → dtype not in filename; official weights are BF16 |
| ernie-image.safetensors (base) | 16-bit | 16.07 GB | Comfy-Org/ERNIE-Image → dtype not in filename; official weights are BF16 |
| ernie-image-turbo-BF16.gguf (turbo) | BF16 | 16.07 GB | unsloth/ERNIE-Image-Turbo-GGUF → |
| ernie-image-turbo-F16.gguf (turbo) | F16 | 16.07 GB | unsloth/ERNIE-Image-Turbo-GGUF → |
| ernie-image-turbo-Q8_0.gguf (turbo) | Q8_0 | 8.69 GB | unsloth/ERNIE-Image-Turbo-GGUF → |
| ernie-image-turbo-Q6_K.gguf (turbo) | Q6_K | 6.79 GB | unsloth/ERNIE-Image-Turbo-GGUF → |
| ernie-image-turbo-Q5_1.gguf (turbo) | Q5_1 | 6.24 GB | unsloth/ERNIE-Image-Turbo-GGUF → |
| ernie-image-turbo-Q5_K_M.gguf (turbo) | Q5_K_M | 5.93 GB | unsloth/ERNIE-Image-Turbo-GGUF → |
| ernie-image-turbo-Q5_K_S.gguf (turbo) | Q5_K_S | 5.86 GB | unsloth/ERNIE-Image-Turbo-GGUF → |
| ernie-image-turbo-Q5_0.gguf (turbo) | Q5_0 | 5.75 GB | unsloth/ERNIE-Image-Turbo-GGUF → |
| ernie-image-turbo-Q4_1.gguf (turbo) | Q4_1 | 5.25 GB | unsloth/ERNIE-Image-Turbo-GGUF → |
| ernie-image-turbo-Q4_K_M.gguf (turbo) | Q4_K_M | 5.02 GB | unsloth/ERNIE-Image-Turbo-GGUF → |
| ernie-image-turbo-Q4_K_S.gguf (turbo) | Q4_K_S | 4.88 GB | unsloth/ERNIE-Image-Turbo-GGUF → |
| ernie-image-turbo-Q4_0.gguf (turbo) | Q4_0 | 4.76 GB | unsloth/ERNIE-Image-Turbo-GGUF → |
| ernie-image-turbo-Q3_K_M.gguf (turbo) | Q3_K_M | 3.91 GB | unsloth/ERNIE-Image-Turbo-GGUF → |
| ernie-image-turbo-Q3_K_S.gguf (turbo) | Q3_K_S | 3.83 GB | unsloth/ERNIE-Image-Turbo-GGUF → |
| ernie-image-turbo-Q2_K.gguf (turbo) | Q2_K | 3.18 GB | unsloth/ERNIE-Image-Turbo-GGUF → |
8B DiT from Baidu (HF API 8.03B BF16), strong text rendering. Turbo = 8 steps (DMD+RL distilled), base = 50 steps. VAE: flux2-vae (336213556, FLUX.2 VAE). Prompt enhancer is an optional LLM that rewrites prompts; skip it to save memory. Native ComfyUI support (Comfy-Org repackage).