In this guide
Short answer: Decode a single still image from MiniMax H3 with the eight-step Turbo LoRA. The listed files total about 43.83 GB on disk. The repositories list MiniMax H3 Community License Agreement.
Output example
This workflow was run on a PC with an RTX 3090. We downloaded the saved output and inspected the image. It is one prompt and seed, not a general quality benchmark.
Prompt used for this example
A cinematic still photograph inside an empty coastal observatory at blue hour. A large brass telescope points through the open circular dome toward a thin crescent moon above a calm sea. The telescope occupies the left foreground; curved dark steel ribs frame the sky overhead, and a narrow warm amber work lamp illuminates the worn wooden floor. Salt haze softens the distant horizon while the brass has engraved rings, tarnished seams and restrained specular highlights. Wide architectural composition with deep navy shadows and warm gold accents. No writing or people.Observed limits: Telescope is tightly cropped at the left and material highlights are somewhat glossy; saved Turbo still recipe only.
The featured image documents the workflow graph. The output example above comes from a separate completed run, with its recorded prompt and limitations.
Download and load the workflow
Drag the downloaded JSON file onto the ComfyUI canvas to load it.
Install the model files and any custom nodes listed below. If the workflow uses input images, video or audio, select your own files in the input nodes before running it.
Which files go where
Place each file under your ComfyUI directory in the folder shown. A text encoder turns the prompt into model conditioning. A VAE decodes the sampled representation into image or audio data. A LoRA is an adapter loaded alongside the base model.
| Part | File | Size | Folder |
|---|---|---|---|
| diffusion_models | minimax_h3_fl2va_pruned_int8_convrot.safetensors | 20.97 GB | models/diffusion_models |
| loras | minimax_h3_fl2v_turbo_8step_v1.0_comfyui_bf16.safetensors | 1.96 GB | models/loras |
| text_encoders | qwen3vl_32b_minimax_h3_nvfp4_awq.safetensors | 15.69 GB | models/text_encoders |
| vae | minimax_h3_video_vae_fp16.safetensors | 5.21 GB | models/vae |
The listed files total about 43.83 GB on disk. Disk size is not a measured VRAM requirement. Reuse matching files already installed rather than downloading a second copy.
Settings in this workflow
| Control | Saved value |
|---|---|
| Steps | 8 |
| Sampler / scheduler | res_multistep / simple |
| Turbo LoRA strength | 1.0 |
| Resolution selector | 16:9, 0.9 megapixels, 8-pixel alignment |
| Empty latent fallback | 1024 × 1024, batch 1 |
| Attention backend | comfy kitchen attention |
How to use it
- Open the text-to-image template and enter your own prompt. No input image is needed.
- Select the four files below, including the eight-step Turbo LoRA.
- Keep the saved eight-step scheduler and choose your aspect ratio in Resolution Selector.
- Generate when ready. VAE Decode produces a still image and Save Image writes it to disk.
What to check before changing the graph
The saved workflow loads the eight-step Turbo LoRA. Its note about 30 default steps and 50 steps refers to the base recipe, so it is not the active setting shown here. To explore the base recipe, bypass the Turbo LoRA and use the corresponding base settings. This still-image graph needs the video VAE but does not load an audio VAE. The notes also list other encoders and four-step adapters; they are alternatives, not additional required downloads.
Licence
The repositories list MiniMax H3 Community License Agreement. This is a community licence with conditions. Read the current terms before commercial use; a public download does not mean unrestricted use. See our model licence tracker for the wider collection.
Troubleshooting
| Problem | Likely cause | Fix |
|---|---|---|
| Missing nodes | Your installation differs from the captured environment | Update ComfyUI and reload the template before reconnecting nodes. |
| Model missing from a dropdown | File is in the wrong folder or the list is stale | Check the exact filename and folder above, then refresh the model list. |
| Unexpected output size or length | A connected input overrides an internal widget | Trace the size and duration connections from the outer graph. |
Where to go next
FAQ
Are these output examples?
The featured image shows the workflow graph. The Output example section shows the actual saved generation from a PC with an RTX 3090.
Can I change the saved settings?
Yes. Keep the documented configuration as a reference and change one control at a time. Samplers, adapters and noise schedules work together, so treat a replacement recipe as a separate configuration.
Sources
- Comfy-Org/MiniMax-H3: model files and repository licence metadata.
Keep building
Get the cheat sheets and planners or browse all ComfyUI workflows.
Need help building a production workflow? Talk to Automate Digital.