fal Launches H3 Max: Post-Trained Video Model Generates 5-Second Clips in Under 3 Seconds
fal has announced H3 Max, a post-trained and inference-optimized derivative of the open-weights MiniMax H3 video model, developed by fal Research in close collaboration with the...