Testing a MiniMax R2V Video Workflow with ChatGPT References

sappigekip · reddit · 2026-08-05

The author shares a video creation test based on the MiniMax R2V model. The workflow involves generating reference images using ChatGPT, running them through the standard MiniMax R2V workflow in ComfyUI, and finally stitching the video clips together in DaVinci Resolve.

While the video generation quality was good, the author noted that the audio automatically generated by the model was not ideal in this particular test case.

Original post →

More from Multimodal

Multimodal channel →