<div data-youtube-video="https://youtu.be/SUJkkT6XLHI"><iframe src="https://www.youtube.com/embed/SUJkkT6XLHI" width="853" height="480" allowfullscreen></iframe></div> 
RunningHub workflow page: https://www.runninghub.ai/zh-cn/post/2086411688710066177?inviteCode=rh-v1111
<h2>MiniMax H3 Reference-To-Video — 24 GB Edition</h2>
A MiniMax H3 video graph from episode 291, part of the six-times-acceleration series that compares model tier, LoRA, attention and cache settings.
<h2>Saved configuration</h2>
27 nodes / 34 links. LoadImage + LoadVideo + LoadAudio. UNET <code>minimax_h3_ref2va_bf16.safetensors</code>; CLIP <code>qwen3vl_32b_minimax_h3_bf16</code>; video fp16/audio fp32 VAEs; turbo-v4 LoRA at 1; euler + beta 6 steps denoise 1.
<h2>How to use</h2>
The 24 GB reference-to-video edition: bf16 ref2va weights and an image + video + audio triple as input. Reference conditioning comes from the still and the clip together; the turbo LoRA keeps the sampler at six steps.
<h2>Limits and not-verified items</h2>
No inference was run for this listing: no speed, VRAM, quality, audio-sync or acceleration-factor claim is made. The "6x" in the episode title is the author's description and was not independently measured. Values are quoted from the saved nodes. H3 weights, LoRAs, encoder and VAEs are not bundled; check their licenses separately. The graph notes credit the official Comfy-Org/MiniMax-H3 model files; preserved as-is, no originality claim. The cover image is a promotional still, not a frame produced by this graph.
Tutorial: <a href="https://www.youtube.com/watch?v=SUJkkT6XLHI" rel="ugc">YouTube</a> · <a href="https://www.bilibili.com/video/BV17nuU6xE3g" rel="ugc">Bilibili</a>. Workflows supplied through AI-KSK; verify workflow, node, model, LoRA and cover rights separately. No blanket commercial or redistribution license is asserted.

MiniMax H3 Reference-to-Video 6x Speedup - 24 GB Edition

<div data-youtube-video="https://youtu.be/SUJkkT6XLHI"></div> 
RunningHub workflow page: https://www.runninghub.ai/zh-cn/post/2086411688710066177?inviteCode=rh-v1111
MiniMax H3 Reference-To-Video — 24 GB Edition
A MiniMax H3 video graph from episode 291, part of the six-times-acceleration series that compares model tier, LoRA, attention and cache settings.
Saved configuration
27 nodes / 34 links. LoadImage + LoadVideo + LoadAudio. UNET <code>minimax_h3_ref2va_bf16.safetensors</code>; CLIP <code>qwen3vl_32b_minimax_h3_bf16</code>; video fp16/audio fp32 VAEs; turbo-v4 LoRA at 1; euler + beta 6 steps denoise 1.
How to use
The 24 GB reference-to-video edition: bf16 ref2va weights and an image + video + audio triple as input. Reference conditioning comes from the still and the clip together; the turbo LoRA keeps the sampler at six steps.
Limits and not-verified items
No inference was run for this listing: no speed, VRAM, quality, audio-sync or acceleration-factor claim is made. The "6x" in the episode title is the author's description and was not independently measured. Values are quoted from the saved nodes. H3 weights, LoRAs, encoder and VAEs are not bundled; check their licenses separately. The graph notes credit the official Comfy-Org/MiniMax-H3 model files; preserved as-is, no originality claim. The cover image is a promotional still, not a frame produced by this graph.
Tutorial: <a href="https://www.youtube.com/watch?v=SUJkkT6XLHI" rel="ugc">YouTube</a> · <a href="https://www.bilibili.com/video/BV17nuU6xE3g" rel="ugc">Bilibili</a>. Workflows supplied through AI-KSK; verify workflow, node, model, LoRA and cover rights separately. No blanket commercial or redistribution license is asserted.

06_MiniMax_H3_____6___24GB___.json

v1.0

https://youtu.be/SUJkkT6XLHI 
 
This workflow is the 24GB general reference-to-video acceleration version for MiniMax H3. It is made for users who need a reference-driven video route while still targeting a lighter memory configuration. 
 
The connected graph uses a MiniMax H3 Ref2VA model route, Qwen3-VL MiniMax text encoding, MiniMax H3 video and audio VAEs, a MiniMaxH3ReferenceToVideo control node, turbo acceleration, and SaveVideo output. The 24GB version is the practical entry point for this reference group. 
 
This package is meant for practical ComfyUI and RunningHub users who want a ready-made MiniMax H3 graph instead of rebuilding loader, encoder, VAE, sampler, and video export connections by hand. The workflow is especially useful for speed comparison, hardware planning, and repeatable prompt tests. Keep the same source image and prompt when comparing versions so the difference you see comes from the workflow route, not from changing creative variables. 
 
Main features: 
 
- MiniMax H3 reference-to-video workflow for 24GB-class use 
- MiniMaxH3ReferenceToVideo node is connected in the active path 
- Reference input helps guide motion, subject, and scene behavior 
- Qwen3-VL MiniMax encoder and MiniMax H3 VAE chain 
- Euler sampling with a beta 6-step schedule in the inspected graph 
- 16:9 output route 
- Good starting point for reference-driven acceleration tests 
 
Suggested workflow: 
 
Use a clean reference image or reference package and keep the prompt explicit about which element should be preserved. Start with simple camera motion and readable action. If the reference relationship is correct, then test the 48GB and 80GB variants with the same source material. 
 
For better comparisons, make one small test first, review identity stability and temporal motion, then raise prompt complexity only after the base route behaves correctly. When using speaking-character prompts, keep the face visible, avoid too many fast cuts, and separate action, expression, and camera instructions clearly. 
 
RunningHub Workflow 
 
Try the workflow online right now - no installation required. 
Workflow: https://www.runninghub.ai/post/2086424087903186946?inviteCode=rh-v1111 
 
If the results meet your expectations, you can later deploy it locally for customization. 
 
Fan Benefits: Register to get 1000 points + daily login 100 points - enjoy 4090 performance and 48 GB super power! 
 
Bilibili Updates (Mainland China &amp; Asia-Pacific) 
 
If you're in the Asia-Pacific region, you can watch the video below to see the workflow demonstration and creative breakdown. 
Bilibili Video: https://www.bilibili.com/video/BV17nuU6xE3g/ 
 
Support Me on Ko-fi 
 
If you find my content helpful and want to support future creations, you can buy me a coffee. 
Every bit of support helps me keep creating. 
Ko-fi: https://ko-fi.com/aiksk 
 
Business Contact 
 
For collaboration or inquiries, please contact aiksk95 on WeChat. 
 
打开下方链接即可在线体验，无需安装。 
工作流：https://www.runninghub.ai/post/2086424087903186946?inviteCode=rh-v1111 
 
如果你觉得效果理想，也可以在本地进行自定义部署。 
 
粉丝福利：注册即可领取 1000 积分，每日登录再领 100 积分，体验 4090 和 48GB 大显存性能。 
 
B站视频（中国大陆及亚太地区） 
 
如果你在中国大陆或亚太地区，可以通过下面的视频查看工作流的实测效果与创作思路。 
B站视频：https://www.bilibili.com/video/BV17nuU6xE3g/ 
 
我会在夸克网盘持续更新模型资源： 
https://pan.quark.cn/s/07bdc81784ce

Files with SHA256: 0d44eb0e7524f82cdfa025b86e88a884cca95472de95a5e605f4c717df616721