Mage-Flow (Microsoft) VRAM requirements
Microsoft's 4B native-resolution model from July 2026, with separate Turbo (4-step) and Edit checkpoints of the same size. Light enough for most GPUs.
TypeImage
Parameters4.12B
Fits entirely from6 GB
8-bit or better from6 GB
01The files, and how much VRAM each needs
| File | Size | Needed | Min. VRAM | Quality | Source |
|---|---|---|---|---|---|
| 16-bit | 8.2 GB | 10.0 GB | 11 GB | the original weights | Comfy-Org/Mage-Flow → |
| INT8 | 4.2 GB | 6.0 GB | 6 GB | practically identical to the original | Comfy-Org/Mage-Flow → |
“Needed” = file + 1 GB working memory + 0.8 GB system reserve.
02By amount of VRAM
03Best GPU for Mage-Flow
The cheapest cards (by launch price) that run it well, and every card sorted by memory: best GPU for Mage-Flow → 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 | Runs well | INT8 | 6.0 GB |
| RTX 3050 6 GB | 6 GB | Runs well | INT8 | 6.0 GB |
| RTX 2070 Super 8 GB | 8 GB | Runs well | INT8 | 6.0 GB |
| RTX 2080 Super 8 GB | 8 GB | Runs well | INT8 | 6.0 GB |
| RTX 3050 8 GB | 8 GB | Runs well | INT8 | 6.0 GB |
| RTX 3060 8 GB | 8 GB | Runs well | INT8 | 6.0 GB |
| RTX 3060 Ti 8 GB | 8 GB | Runs well | INT8 | 6.0 GB |
| RTX 3070 8 GB | 8 GB | Runs well | INT8 | 6.0 GB |
| RTX 3070 Ti 8 GB | 8 GB | Runs well | INT8 | 6.0 GB |
| RTX 4060 8 GB | 8 GB | Runs well | INT8 | 6.0 GB |
| RTX 4060 Ti 8 GB | 8 GB | Runs well | INT8 | 6.0 GB |
| RTX 5050 8 GB | 8 GB | Runs well | INT8 | 6.0 GB |
| RTX 5060 8 GB | 8 GB | Runs well | INT8 | 6.0 GB |
| RTX 5060 Ti 8 GB | 8 GB | Runs well | INT8 | 6.0 GB |
| RX 7600 8 GB | 8 GB | Runs well | INT8 | 6.0 GB |
| RX 9050 8 GB | 8 GB | Runs well | INT8 | 6.0 GB |
| RX 9060 XT 8 GB | 8 GB | Runs well | INT8 | 6.0 GB |
| Arc B570 10 GB | 10 GB | Runs well | INT8 | 6.0 GB |
| RTX 3080 10 GB | 10 GB | Runs well | INT8 | 6.0 GB |
| RTX 2080 Ti 11 GB | 11 GB | Runs well | 16-bit | 10.0 GB |
| Arc B580 12 GB | 12 GB | Runs well | 16-bit | 10.0 GB |
| RTX 2060 12 GB | 12 GB | Runs well | 16-bit | 10.0 GB |
| RTX 3060 12 GB | 12 GB | Runs well | 16-bit | 10.0 GB |
| RTX 3080 12 GB | 12 GB | Runs well | 16-bit | 10.0 GB |
| RTX 3080 Ti 12 GB | 12 GB | Runs well | 16-bit | 10.0 GB |
| RTX 4070 12 GB | 12 GB | Runs well | 16-bit | 10.0 GB |
| RTX 4070 Super 12 GB | 12 GB | Runs well | 16-bit | 10.0 GB |
| RTX 4070 Ti 12 GB | 12 GB | Runs well | 16-bit | 10.0 GB |
| RTX 5070 12 GB | 12 GB | Runs well | 16-bit | 10.0 GB |
| RX 7700 XT 12 GB | 12 GB | Runs well | 16-bit | 10.0 GB |
| RX 9070 GRE 12 GB | 12 GB | Runs well | 16-bit | 10.0 GB |
| Arc A770 16 GB | 16 GB | Runs well | 16-bit | 10.0 GB |
| RTX 4060 Ti 16 GB | 16 GB | Runs well | 16-bit | 10.0 GB |
| RTX 4070 Ti Super 16 GB | 16 GB | Runs well | 16-bit | 10.0 GB |
| RTX 4080 16 GB | 16 GB | Runs well | 16-bit | 10.0 GB |
| RTX 4080 Super 16 GB | 16 GB | Runs well | 16-bit | 10.0 GB |
| RTX 5060 Ti 16 GB | 16 GB | Runs well | 16-bit | 10.0 GB |
| RTX 5070 Ti 16 GB | 16 GB | Runs well | 16-bit | 10.0 GB |
| RTX 5080 16 GB | 16 GB | Runs well | 16-bit | 10.0 GB |
| RX 7600 XT 16 GB | 16 GB | Runs well | 16-bit | 10.0 GB |
| RX 7800 XT 16 GB | 16 GB | Runs well | 16-bit | 10.0 GB |
| RX 7900 GRE 16 GB | 16 GB | Runs well | 16-bit | 10.0 GB |
| RX 9060 XT 16 GB | 16 GB | Runs well | 16-bit | 10.0 GB |
| RX 9070 16 GB | 16 GB | Runs well | 16-bit | 10.0 GB |
| RX 9070 XT 16 GB | 16 GB | Runs well | 16-bit | 10.0 GB |
| RX 7900 XT 20 GB | 20 GB | Runs well | 16-bit | 10.0 GB |
| Arc Pro B60 24 GB | 24 GB | Runs well | 16-bit | 10.0 GB |
| RTX 3090 24 GB | 24 GB | Runs well | 16-bit | 10.0 GB |
| RTX 3090 Ti 24 GB | 24 GB | Runs well | 16-bit | 10.0 GB |
| RTX 4090 24 GB | 24 GB | Runs well | 16-bit | 10.0 GB |
| RX 7900 XTX 24 GB | 24 GB | Runs well | 16-bit | 10.0 GB |
| Arc Pro B70 32 GB | 32 GB | Runs well | 16-bit | 10.0 GB |
| RTX 5090 32 GB | 32 GB | Runs well | 16-bit | 10.0 GB |
| Laptop GPUs | ||||
| RTX 3050 Laptop 4 GB | 4 GB | Offload only | INT8 | 6.0 GB |
| RTX 3050 Ti Laptop 4 GB | 4 GB | Offload only | INT8 | 6.0 GB |
| RTX 2060 Laptop 6 GB | 6 GB | Runs well | INT8 | 6.0 GB |
| RTX 3050 Laptop 6 GB | 6 GB | Runs well | INT8 | 6.0 GB |
| RTX 3060 Laptop 6 GB | 6 GB | Runs well | INT8 | 6.0 GB |
| RTX 4050 Laptop 6 GB | 6 GB | Runs well | INT8 | 6.0 GB |
| RTX 2070 Laptop 8 GB | 8 GB | Runs well | INT8 | 6.0 GB |
| RTX 2070 Super Laptop 8 GB | 8 GB | Runs well | INT8 | 6.0 GB |
| RTX 2080 Laptop 8 GB | 8 GB | Runs well | INT8 | 6.0 GB |
| RTX 2080 Super Laptop 8 GB | 8 GB | Runs well | INT8 | 6.0 GB |
| RTX 3070 Laptop 8 GB | 8 GB | Runs well | INT8 | 6.0 GB |
| RTX 3070 Ti Laptop 8 GB | 8 GB | Runs well | INT8 | 6.0 GB |
| RTX 3080 Laptop 8 GB | 8 GB | Runs well | INT8 | 6.0 GB |
| RTX 4060 Laptop 8 GB | 8 GB | Runs well | INT8 | 6.0 GB |
| RTX 4070 Laptop 8 GB | 8 GB | Runs well | INT8 | 6.0 GB |
| RTX 5050 Laptop 8 GB | 8 GB | Runs well | INT8 | 6.0 GB |
| RTX 5060 Laptop 8 GB | 8 GB | Runs well | INT8 | 6.0 GB |
| RTX 5070 Laptop 8 GB | 8 GB | Runs well | INT8 | 6.0 GB |
| RX 7600M 8 GB | 8 GB | Runs well | INT8 | 6.0 GB |
| RX 7600M XT 8 GB | 8 GB | Runs well | INT8 | 6.0 GB |
| RX 7600S 8 GB | 8 GB | Runs well | INT8 | 6.0 GB |
| RX 7700S 8 GB | 8 GB | Runs well | INT8 | 6.0 GB |
| RTX 4080 Laptop 12 GB | 12 GB | Runs well | 16-bit | 10.0 GB |
| RTX 5070 Laptop 12 GB | 12 GB | Runs well | 16-bit | 10.0 GB |
| RTX 5070 Ti Laptop 12 GB | 12 GB | Runs well | 16-bit | 10.0 GB |
| RX 7800M 12 GB | 12 GB | Runs well | 16-bit | 10.0 GB |
| RTX 3080 Laptop 16 GB | 16 GB | Runs well | 16-bit | 10.0 GB |
| RTX 3080 Ti Laptop 16 GB | 16 GB | Runs well | 16-bit | 10.0 GB |
| RTX 4090 Laptop 16 GB | 16 GB | Runs well | 16-bit | 10.0 GB |
| RTX 5080 Laptop 16 GB | 16 GB | Runs well | 16-bit | 10.0 GB |
| RX 7900M 16 GB | 16 GB | Runs well | 16-bit | 10.0 GB |
| RTX 5090 Laptop 24 GB | 24 GB | Runs well | 16-bit | 10.0 GB |
| Unified memory | ||||
| Radeon 8060S (Strix Halo) 96 GB | 96 GB | Runs well | 16-bit | 10.0 GB |
| Radeon 8065S (Gorgon Halo) 160 GB | 160 GB | Runs well | 16-bit | 10.0 GB |
| Apple Silicon Macs (by memory) | ||||
| Mac 16 GB | 12.7 GB | Runs well | 16-bit | 10.0 GB |
| Mac 18 GB | 14.4 GB | Runs well | 16-bit | 10.0 GB |
| Mac 24 GB | 19.6 GB | Runs well | 16-bit | 10.0 GB |
| Mac 32 GB | 26.8 GB | Runs well | 16-bit | 10.0 GB |
| Mac 36 GB | 30.2 GB | Runs well | 16-bit | 10.0 GB |
| Mac 48 GB | 40.2 GB | Runs well | 16-bit | 10.0 GB |
| Mac 64 GB | 55.7 GB | Runs well | 16-bit | 10.0 GB |
| Mac 96 GB | 85 GB | Runs well | 16-bit | 10.0 GB |
| Mac 128 GB | 115.4 GB | Runs well | 16-bit | 10.0 GB |
| Mac 192 GB | 175.4 GB | Runs well | 16-bit | 10.0 GB |
| Mac 256 GB | 236.9 GB | Runs well | 16-bit | 10.0 GB |
| Mac 512 GB | 498.1 GB | Runs well | 16-bit | 10.0 GB |
05Text encoder, VAE and other files
| File | Folder | Size | When | |
|---|---|---|---|---|
| Qwen3-VL 4B BF16 qwen3vl_4b_bf16.safetensors | models/text_encoders | 8.9 GB | required | Download → |
| Mage-Flow VAE mage_flow_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).
06Where the files go in ComfyUI
| File | Folder | Loader node |
|---|---|---|
| Diffusion model (.safetensors: 16-bit, FP8, INT8) | ComfyUI/models/diffusion_models | Load Diffusion Model |
| 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 Mage-Flow
No trainer documentation with a VRAM figure for this model was found yet. What the trainers say for other models →
09Every file tracked for Mage-Flow
| File | Type | Size | Repo |
|---|---|---|---|
| mage_flow_bf16.safetensors (rl) | BF16 | 8.23 GB | Comfy-Org/Mage-Flow → |
| mage_flow_base_bf16.safetensors (base) | BF16 | 8.23 GB | Comfy-Org/Mage-Flow → |
| mage_flow_edit_bf16.safetensors (edit) | BF16 | 8.23 GB | Comfy-Org/Mage-Flow → |
| mage_flow_edit_base_bf16.safetensors (edit-base) | BF16 | 8.23 GB | Comfy-Org/Mage-Flow → |
| mage_flow_turbo_bf16.safetensors (turbo) | BF16 | 8.23 GB | Comfy-Org/Mage-Flow → |
| mage_flow_edit_turbo_bf16.safetensors (edit-turbo) | BF16 | 8.23 GB | Comfy-Org/Mage-Flow → |
| mage_flow_int8_convrot.safetensors (rl) | INT8 | 4.16 GB | Comfy-Org/Mage-Flow → |
| mage_flow_edit_int8_convrot.safetensors (edit) | INT8 | 4.16 GB | Comfy-Org/Mage-Flow → |
| mage_flow_turbo_int8_convrot.safetensors (turbo) | INT8 | 4.16 GB | Comfy-Org/Mage-Flow → |
| mage_flow_edit_turbo_int8_convrot.safetensors (edit-turbo) | INT8 | 4.16 GB | Comfy-Org/Mage-Flow → |
| mageflow-edit-turbo-nvfp4.gguf (edit-turbo) | NVFP4 | 2.37 GB | gguf-org/mageflow-gguf → NVFP4 packed in GGUF container (gguf-org / calcuis 'gguf' node); only quant available, no K-quants |
4.1B NR-MMDiT (HF API 4,115,745,408 BF16), native-resolution rectified flow. Plain 'mage_flow' = RL-aligned (20 steps), base = 30 steps, turbo = 4 steps; edit/edit-base/edit-turbo are separate instruction-edit checkpoints. VAE: mage_flow_vae_bf16 (345053056), own one-step Mage-VAE. Native ComfyUI support (docs.comfy.org workflow). Official weights under mage-flow-community org.