Hands-on with Seedance and MiniMax: prompt structure no longer matters much

apostraphi · x · 2026-09-22

The author shares experiments with Seedance and MiniMax for video generation: no need to obsess over prompt structure — the models already have that context, so you focus only on the details of the shot. Wrong mood, red sky? Swap the frame and go again. The thread also compares how the two models handle grain, skin textures, prompt adherence, and image references.

Original post →

More from Multimodal

Multimodal channel →