fal releases H3 Max, claiming top video benchmark rankings and 3-second generation for 5-second clips
- fal released H3 Max, a post-trained version of the open-weights MiniMax H3 video model, for text-to-video and image-to-video generation through its API, Playground, and Agent.
- fal says H3 Max generates a five-second video in about three seconds, about 35 times the throughput of the official MiniMax H3 endpoint and 15 times faster on average than comparably rated models in its tests.
- Design Arena ranks H3 Max first on its recent image-to-video leaderboard at 1,341 Elo, ahead of official MiniMax H3 at 1,333.
- Artificial Analysis ranks fal's H3 model first on its image-to-video-with-audio leaderboard, with 1,201 Elo across 2,177 samples.
- fal post-trained the model with new data and an in-house reinforcement learning framework, while co-designing inference optimizations that it retained only when human preference evaluations preserved quality gains; launch pricing is $0.04 per second at 768p until September 7, then $0.08 per second.