Paper 2607.18703
Generative World Renderer at the Speed of Play
- Published
- Jul 2026
- Research lab
- Independent
- Citations
- 0
- GitHub
- 53 stars
01 In brief
Summary
AlayaRenderer-Flash is a real-time generative world renderer that accelerates the offline AlayaRenderer from 0.56 FPS to 31.54 FPS, enabling interactive, prompt-controllable gameplay.
It reformulates the original renderer into a few-step autoregressive streaming model with three key improvements: autoregressive generation over unbounded G-buffer streams, distillation of the 50-step denoising schedule into 4 steps, and lightweight distilled codecs for efficient latent encoding and frame reconstruction.
The model retains the teacher's G-buffer and text-prompt interfaces, allowing continuous rendering with seamless prompt switching.
Evaluated on a Black Myth: Wukong dataset, AlayaRenderer-Flash achieves the best trade-off between quality and efficiency compared to baselines like RGB↔X and FrameDiffuser, supporting causal streaming, prompt control, and real-time inference.
Integrated with the SuperTuxKart engine, it sustains 30 FPS during live gameplay, demonstrating a fully playable generative world.
02 From the paper
Abstract
Generative world renderer AlayaRenderer receives structured world states exported from physics engines and synthesizes RGB frames. Unlike models that generate frames from text/control-hints prompts, AlayaRenderer preserves scene structure without altering the underlying world dynamics. This demonstrates an alternative path toward interactive world modeling and user-controllable play. However, the original AlayaRenderer is too computationally expensive for real-time deployment. This technical report introduces AlayaRenderer-Flash, a real-time-oriented generative forward world renderer that pushes AlayaRenderer from 0.56 FPS to 31.54 FPS, reaching the speed of play. AlayaRenderer-Flash reformulates the original renderer as a few-step autoregressive streaming model and introduces lightweight distilled codecs for efficient latent encoding and frame reconstruction. It retains the teacher model's G-buffer and text-prompt interfaces while enabling continuous rendering over input streams of unbounded length. We evaluate AlayaRenderer-Flash on G-buffer streams across content preservation, temporal consistency, cross-window stability, prompt controllability, and runtime efficiency. Our results show that AlayaRenderer-Flash substantially reduces inference cost while preserving the core rendering capabilities of the teacher model. By integrating AlayaRenderer-Flash with a physics engine, we build a fully playable generative world running at 30 FPS.