MiniMax H3 Max text-to-video API launches on fal.ai

venturetwins · x · 2026-08-27

MiniMax H3 Max text-to-video model is now available via API on fal.ai. As a post-trained variant of H3, it is tuned for stronger prompt adherence and better aesthetics, co-optimized with a custom inference stack for higher throughput. Examples demonstrate complex post-apocalyptic sci-fi scene generation, supporting customizable duration (768P resolution), aspect ratio, and prompt expansion modes.

Related event: fal's H3 Max Topples Video Leaderboard With Massive Speed Gains(13 posts)→

Original post →

More from Multimodal

Multimodal channel →