Find your next building block

Direct an ongoing generated video stream with changing prompts through fal's experimental realtime API.

H3 Max is the speed-optimized build of MiniMax H3, the open-weight multimodal video model. It turns a prompt or an image into a 5-second 768p clip in roughly three seconds — faster than the video plays. h3max.info collects the official links, APIs, open weights, guides and live projects so you can start building in minutes.

Run a released 14B long-video model with training code, inference tools and distilled checkpoints.

A preview audio-visual world model that responds to navigation actions while generating synchronized media.

An endless AI-generated television channel where viewers write what should air next.

Generate multi-view-consistent frames from a reference image and target camera poses on consumer-oriented hardware.

Two released research systems in one repository: a long-horizon multi-shot audio-video generator and an interactive omnimodal world model.

Real-time instruction-guided editing of open-ended video streams with an autoregressive diffusion transformer.

Explore a controllable video world model with released base and fast inference variants.

Edit a streaming video with diffusion while reusing unchanged regions and extending to longer sequences.

An interactive video world model that explicitly simulates dynamics outside the current camera view.

Explore long-running autoregressive video generation and the infrastructure behind interactive streams.

A streaming interactive world-model release focused on 720p generation and long-horizon memory.

A playable real-time world and human action model available as an experimental Copilot Labs experience.

A Minecraft-specific interactive video world model with downloadable checkpoints and a local web demo.

A hosted text-to-video and image-to-video model with 480P, 768P, 2K, and 4K price records.

A hosted MiniMax H3 variant with text-to-video and image-to-video endpoints at 480P and 768P price tiers.

A hosted text-to-video and image-to-video model record focused on fast clip generation and source-linked cost planning.

Learn and reproduce the full pipeline for converting video diffusion backbones into interactive world models.

A hosted real-time conversational video model for adding live visual agents to applications.

A public research preview of a narrow-domain playable world model that streams newly generated frames.

A world-model experience for creating and playing generated games through words, voice, and in-world actions.

A hosted multimodal world model for continuous interactive video, audio and state generation.

A developer platform for deploying and streaming real-time generative-media and world-model sessions.