Fal's H3 Max Live generates video faster than you can watch it, at real-time factor 1 and 768p

Latent Space (AINews)·medium signalXBlueskyLinkedInCopy link

AINews reported September 1 that Fal launched fal.live around H3 Max Live, an autoregressive continuous video generation mode built on MiniMax's H3 and post-trained for cost and quality, reaching a real-time factor of about 1 at 768p with a claimed 35x speed improvement over the official endpoints through Fal's inference stack. H3 Max Director holds up to two minutes of context for continuous generation, and the platform layers audience steering on top: viewers vote, an LLM turns votes into prompts, and the prompt appears on screen within seconds. Crossing real-time factor 1 is the threshold that turns video generation from a render job into a stream.