MiniMax H3, a video model, makes surprisingly good 4–16 MP stills in 6 steps

ReferenceFluffy6159 · reddit · 2026-11-17

A free ComfyUI workflow (H3 Image AIO) shows MiniMax H3, a video model, doubling as a text-to-image and editing tool, with backgrounds showing real depth instead of dreamy blur.

Key takeaways from testing:

The workflow supports up to 6 reference images/video, a reserved trigger-word box, color adjustment, and 4 example prompts. Author's score: Qwen Image 2.1 remains the daily driver (9/10); H3 Image 8.5/10 with text rendering close to MOSS Image.

Original post →

More from Multimodal

Multimodal channel →