# A quick Minimax H3 news round-up - 3rd October 2026 — retained record

> Retained reddit record from reddit:comfyui.

Canonical: https://highsignal.app/data/records/896762689901c6e0

[← sources](https://highsignal.app/data)


source: [reddit](https://highsignal.app/data/reddit) (reddit:comfyui)  published 2026-10-03 13:41 ingested 2026-10-04 02:38

[original source ↗](https://www.reddit.com/r/comfyui/comments/1wwnkof/a_quick_minimax_h3_news_roundup_3rd_october_2026/)

## retained text



```
Another quick Minimax H3 news and goodies round-up, for those who may have missed some items. 

-> I've organised my Reddit posts into a basic themed H3 directory page on the Web. It doesn't aim to be completist, but it should be useful for quick lookups.

https://jurn.link/dazposer/index.php/h3/

-> A basic Refmod creator workflow for ComfyUI.

https://civitai.com/models/2979689/basic-refmod-creator-minimax-h3

-> A workflow to do character replacement, then it does... "an independent face refine pass over the generated clip to clean up facial detail and identity stability. [This] detects faces with YuNet (ONNX face detection), crops a 512 AV latent window around the tracked face, re-samples it with a separate *minimax_h3_fl2va_int8_convrot* model using a low-denoise 12-step schedule (0.45 denoise, dual_clock_euler / native_flow), and stitches the refined crop back into the full frame."

https://civitai.com/models/2980013/minimax-h3-role-swap-face-refine

-> Another motion combat reasoning LoRA, which the samples suggest it might be especially useful for mediaeval swordfighting?

https://civitai.com/models/2979777/minimax-h3-motion-combat-reasoning-or-ref2va-or-fl2va-or-hybrid

-> A LoRA that aims to boost the realism of an amateur 'digicam recording' style.

https://huggingface.co/RunningHubAI/rh-minimax-h3-digicam-realism-improve-lora

-> A new Multi-language Subtitle Generator. The speech is transcribed from the video by local models and translated by a local LLM, and written to local .SRT subtitles files.

https://github.com/dodgypast/mlsubgen

-> A Claude Code 'Skill' to expand short-code camera prompts into full-blown subject-aware Minimax H3 camera prompts... e.g. you start with "/MS /LOWANGLE /DOLLYIN:MS>MCU:SLOW" and then it's expanded.

https://github.com/SK-SL-source/cinematic-director-camera-dsl

->  And finally, a replacement stack to run Minimax H3 faster on a Intel Arc Pro B70 graphics card. Made with C++, Rust, Typescript.

https://github.com/maxfridbe/sycl_minimax_h3

~ OLD POSTS ~

https://old.reddit.com/r/comfyui/comments/1ww58xn/a_quick_minimax_h3_news_roundup_2nd_october_2026/

https://old.reddit.com/r/comfyui/comments/1wu8lkt/a_quick_minimax_h3_news_roundup_30th_september/

https://old.reddit.com/r/comfyui/comments/1wtgns0/a_quick_minimax_h3_news_roundup_29th_september/

https://old.reddit.com/r/comfyui/comments/1wsmjq1/a_quick_minimax_h3_news_roundup_28th_september/

https://old.reddit.com/r/comfyui/comments/1wrqm3l/a_quick_minimax_h3_news_roundup_27th_september/

https://old.reddit.com/r/comfyui/comments/1wqwpah/a_quick_minimax_h3_news_roundup_26th_september/ (See 26th September post, for links to older posts)

https://old.reddit.com/r/comfyui/comments/1wgc4lj/a_quick_minimax_h3_news_roundup_15th_september/ (See 15th September post, for links to even older posts)

https://old.reddit.com/r/comfyui/comments/1w5i9iq/a_quick_minimax_h3_news_roundup_2nd_september_2026/ (See 2nd September post, for links to the starting posts)
```



document fetched 2026-10-04 00:35 · canonical https://reddit.com/r/comfyui/comments/1wwnkof/a_quick_minimax_h3_news_roundup_3rd_october_2026
