What kind of model is Happy Oyster?

A real-time interactive 3D world model from Alibaba's ATH Innovation Division. It generates explorable environments rather than passive video.

When can I use Happy Oyster?

Limited early access opened April 16, 2026. General availability has not been announced.

How is Happy Oyster different from Sora or Veo?

Sora and Veo produce video clips you watch. Happy Oyster produces 3D worlds you can re-enter and explore from new camera angles.

Is there a public API?

An API has not been publicly documented at the time of writing.

Does Happy Oyster generate audio?

Yes, audio is co-generated natively with the visual scene, so footsteps and ambience match the world without post-processing.

Alibaba ATH Innovation Division vs Runway AI, Inc.

Happy Oyster vs Runway Gen-4.5

Happy Oyster (Real-time interactive 3D world model) compared to Runway Gen-4.5 (A new frontier for video and world simulation.).

Happy Oyster and Runway Gen-4.5 target adjacent jobs but take different approaches. This page compares them side by side on output paradigm, access, capabilities, and positioning — based on vendor-stated claims as of 2026-04-20 / 2026-04-21.

At a Glance

Alibaba ATH Innovation Division

Happy Oyster

Real-time interactive 3D world model

Re-explorable 3D output instead of one-shot video
Two distinct authoring modes (Directing + Wandering)
Native audio-video co-generation in a world model

See Happy Oyster details →

Runway AI, Inc.

Runway Gen-4.5

A new frontier for video and world simulation.

Built on GWM-1, an advanced General World Model architecture that intrinsically understands and simulates real-world physics.
Offers infinite character and object consistency across multiple angles and scenes using only a single reference image.
Provides an integrated node-based workflow system that lets users chain multiple AI models and modalities directly within the web interface.

See Runway Gen-4.5 details →

How They Compare

Dimension	Happy Oyster	Runway Gen-4.5
Modality	3d-world, interactive, audio-video	text-to-video, image-to-video, video-to-video, text-to-audio
Release status	early-access (2026-04-16)	ga (2025-12-11)
Capabilities	Directing Mode · Wandering Mode · Native Audio-Video Co-generation · 3D World Generation	Character Consistency · Physics Simulation · Cinematic Camera Control · Video to Video
Output type	Interactive 3D world (not pre-rendered video)	—
Modes	Directing + Wandering	—
Audio	Natively co-generated with visuals	—
Access	Limited early access (April 2026)	—
API availability	Not publicly documented	—
Pricing	Not announced	—
Game-engine export	Not confirmed	—
Max Resolution	—	4K (via built-in upscale)
Max Duration	—	Up to 34 seconds (via extend tool)
Architecture	—	GWM-1 (General World Model)

Which Should You Choose?

Pick Happy Oyster if you need: Re-explorable 3D output instead of one-shot video.
Pick Runway Gen-4.5 if you need: Built on GWM-1, an advanced General World Model architecture that intrinsically understands and simulates real-world physics..
Both come from different vendors — consider your existing stack.