Tencent's Wizard World Model Enables Real-Time Agentic Control of Any Game World
tencent · hf · 2026-11-02
Tencent with NUS and NTU introduces Wizard, an agentic game world model with real-time, precise character control.
- Pain point: Existing video world models only support simple game settings and predefined motions, generalize poorly to unseen environments, and lack real-time interactivity.
- Method: A dataset spanning 750 scenes and 178 characters/actions uses skeleton and geometry info as structured motion proxies to condition a video generation model, plus an adaptive masking mechanism improving alignment between conditioning signals and generated rollouts.
- Agentic framework: Game states are explicitly maintained and updated, letting players configure worlds and exercise fine-grained motion control with zero-shot generalization to novel scenes and characters.
More from Multimodal
- Akatz Labs releases open-weight Krea 2 OpenCast LoRA for stylized character creation — PurzBeats · 2026-11-02
- 'One Breath': AI-Generated Short Film Wows Reddit With Near-Photographic Realism — Pnarpok · 2026-11-02
- Redditor experiments with AI song covers using open-source YuE2 — AndalusianGod · 2026-11-02
- Are the default Minimax H3 workflows good enough? Unconnected nodes spark doubts — dtdisapointingresult · 2026-11-02
- 'Dancer 5': a striking AI-generated dance video demo — No-While1332 · 2026-11-02
- FrameForge Reuses Video Model Priors for Unified, Adaptive Conditional Image Generation — FrameXAI · 2026-11-02