ComfyUI Node
MiniMax H3 Prompt (Qwen) Multimodal
A ComfyUI node in MiniMax H3 with 23 inputs and 2 outputs.
MiniMax H3 Prompt (Qwen) Multimodal
- image1
- image2
- image3
- image4
- video
- positive_prompt
- negative_prompt
◄model▾►
◄mmprojNone►
◄chat_handlerQwen3.5►
◄n_ctx8192►
◄user_prompt►
◄modeAuto►
◄stylecinematic►
◄duration10►
◄aspect_ratio16:9►
◄include_audiotrue►
◄seed0►
◄max_tokens2048►
◄temperature0.80►
◄top_p0.90►
◄top_k30►
◄max_frames16►
◄max_size512►
◄custom_system_prompt►
CategoryMiniMax H3
Inputs (23)
| Name | Type | Default | Description |
|---|---|---|---|
| model | COMBO | 1 options: （未找到模型，请检查 models/LLM 目录） | |
| mmproj | COMBO | None | 1 options: None |
| chat_handler | COMBO | Qwen3.5 | 20 options: None, LLaVA-1.5, LLaVA-1.6, Moondream2, nanoLLaVA, llama3-Vision-Alpha, +14 |
| n_ctx | INT | 81921024–327680 | 上下文长度限制 |
| user_prompt | STRING | — | |
| mode | COMBO | Auto | Auto=按图片数量自动判断(1张=单图分析, 2+张=多图故事推理), I2VA=首帧图生视频, FL2VA=首尾帧生视频, L2VA=尾帧生视频, Ref2VA=全参考生视频 |
| style | COMBO | cinematic | 8 options: cinematic, live-action, 2D-animated, 3D CG, claymation, watercolor, +2 |
| duration | INT | 104–15 | 视频时长（秒），MiniMax H3 支持 4-15 秒 |
| aspect_ratio | COMBO | 16:9 | 3 options: 16:9, 9:16, 1:1 |
| include_audio | BOOLEAN | true | 是否在提示词中包含环境音和 BGM 描述 |
| seed | INT | 00–18446744073709550000 | — |
| max_tokens | INT | 20480–8192 | — |
| temperature | FLOAT | 0.800–2 | — |
| top_p | FLOAT | 0.900–1 | — |
| top_k | INT | 300–1000 | — |
| max_frames | INT | 162–128 | 视频参考采样帧数 |
| max_size | INT | 512128–4096 | 图片/视频帧最大边长（限制 token 占用） |
| image1opt | IMAGE | — | |
| image2opt | IMAGE | — | |
| image3opt | IMAGE | — | |
| image4opt | IMAGE | — | |
| videoopt | IMAGE | — | |
| custom_system_promptopt | STRING | — |
Outputs (2)
| Name | Type | Description |
|---|---|---|
| positive_prompt | STRING | — |
| negative_prompt | STRING | — |