NVIDIA research model that generates real-time, action-controlled multi-agent video of shared worlds, from two to four players.
Updated EU status checked Last shown
Sources: the vendor's own pages (linked above) and creator videos at the linked timestamps. EU status is information, not legal advice; see how we check.
γ-World (Gamma World) is a generative multi-agent world model from NVIDIA research. It takes a separate action stream for each agent and produces a coherent, shared multi-view video rollout, so several players, robots or embodied agents can act independently in the same evolving environment.
The model introduces Simplex Rotary Agent Encoding, a parameter-free extension of 3D RoPE that places agents at the vertices of a regular simplex in rotary angle space. Every agent gets a distinct phase and all agents are permutation-equivalent, so no learned per-slot identities or fixed agent ordering are needed. Sparse Hub Attention uses learnable hub tokens to mediate cross-agent communication, which makes attention cost linear rather than quadratic in the number of agents. A bidirectional multi-agent teacher is distilled into a block-causal student that uses KV caching for streaming at 24 FPS.
Watch the moment: the exact second top creators show it. Videos load only when you press play, via youtube-nocookie.com.
Multi-agent world simulator that generates real-time 24 fps video of two to four players in a shared environment, using simplex rotary agent encoding. Code, training scripts and dataset tool are coming soon.
Closest neighbours by meaning, not just by category.
Model · NVIDIA
NVIDIA's open omni-model family for physical AI: world simulation, reasoning and action generation for robots and AVs.
Model
Research multi-player video world model with hierarchical communication, trained on 4 players and run with up to 10.
Model · Skywork
Real-time playable world model from Skywork pairing a game engine and coding agent with a few-step causal video model.
Model · Nvidia
Nvidia research model that understands and generates images and video directly in pixel space, with no VAE or vision encoder.
Model · NVIDIA
NVIDIA research system where a code game engine drives a two-step streaming MiniMax-H3 video model for real-time rendering.
Model · NVIDIA
NVIDIA's real-time generative world model for closed-loop autonomous vehicle simulation, built on Cosmos.