cs.CVNov 19, 2026

AURA: AUdio-dRiven streaming Avatar

Authors: Nathaniel Cohen, Nicolas Dufour, Amélie Royer, Alasdair Newson, Patrick Pérez

Organizations: Kyutai Paris, France · ISIR, Sorbonne Université Paris, France · Google DeepMind (work done while at Kyutai)

Abstract

Current audio-driven video diffusion models can synthesize realistic talking-head videos but generally require dozens of diffusion denoising steps over multi-second video chunks, making them unsuitable for streaming applications such as voice assistants and virtual avatars, where low-latency and online causal generation are required properties. AURA generates video chunk by chunk with a rolling KV cache, requiring only the audio of the current chunk to start generating, rather than a multi-second audio look-ahead like bidirectional models. It comes in two sizes: AURA 1.3B streams 576p video in real time on a single GPU (20.7 fps for a 16 Hz playback rate), making it suited to interactive avatar applications, and AURA 5B streams 720p video at 9.7 fps. Our two-stage distillation pipeline first aligns the causal student with the teacher, by training it to match the teacher's trajectories with fewer diffusion steps and a chunk-wise sliding window causal attention mask, then applies Distribution Matching Distillation (DMD) with backward simulation to close the remaining quality gap. On standard talking-head benchmarks, AURA 1.3B runs 23x faster than its own bidirectional teacher and reaches a quality comparable to TalkVerse at a lower resolution, and AURA 5B matches its TalkVerse teacher at 720p while running 17x faster. To support reproducibility and further research on streaming avatar generation, we release the weights of both models along with a dataset of 30K synthetic videos and their precomputed teacher trajectories.

Figures & tables

Appendix figures & tables11 assets

Supplementary material from the paper’s appendix.

Appendix

Explore similar work

CardsList
  1. AvatarForcing: One-Step Streaming Talking Avatars via Local-Future Sliding-Window Denoising

    Mar 15, 2026Liyuan Cui, Wentao Hu, Wenyuan Zhang +3Audio-Video GenerationVideo Diffusion Models

  2. Vorch-Streamer: Extending Human Audio-Visual Generation to Real-Time Long-Form Streaming

    Aug 6, 2026Menglin Han, Yang Ding, Yulei Lu +6T2V GenerationLong-Horizon Video Generation