Alibaba's Wan-Streamer v0.2 Triples Video Resolution Without Adding Latency
Alibaba's Wan-Streamer v0.2 runs full-duplex video conversations at 640×368 and 25 FPS with ~550ms total latency, all inside a single Transformer
- Wan-Streamer v0.2 is a single Transformer model that conducts full-duplex video conversations with ~550ms total latency.
- Resolution jumps from 192×336 (v0.1) to 640×368 at 25 FPS, with ~200ms model-side latency preserved.
- A new Thinker–Performer split lets multi-GPU Ulysses parallelism scale visual quality without increasing interaction delay.
- Unlike cascaded pipelines, all perception, reasoning, speech, and video generation are learned jointly in one model.
- The model supports arbitrary personas (humans, pets, anime) described in natural language, not fixed avatar rigs.
- No open weights or public API yet; available as paper and demo only.
Wan-Streamer v0.2 is Alibaba's latest research push toward something that has been surprisingly hard to build: a single AI model that can simultaneously watch you, listen to you, think, and respond with synchronized speech and video , in real time. The new version ships a major visual upgrade without touching the latency, going from a blurry 192p proof-of-concept to a crisp 640×368 stream at 25 FPS, all while keeping the total round-trip time at roughly 550ms.
One model to rule all the modalities
Most "real-time" AI video systems are actually pipelines: a speech recognizer feeds a language model, which feeds a text-to-speech engine, which drives a separate avatar renderer. Each handoff adds latency and a new failure mode. Unlike cascaded interactive systems that rely on separate VAD, ASR, language, TTS, audio-driven animation, or video-generation modules, Wan-Streamer does not rely on external modules , perception, reasoning, generation, response timing, turn management, and cross-modal synchronization are learned jointly within one unified model.
Wan-Streamer is a native-streaming, end-to-end interactive foundation model designed from the ground up for real-time, low-latency, full-duplex audio-visual interaction. It seamlessly models language, audio, and video as both input and output within a single Transformer, where the sequence is represented as interleaved visual, audio, and text tokens, coordinated by block-causal attention for incremental streaming. Block-causal attention means the model can look back at everything it has already generated, but cannot peek ahead , a constraint that makes true streaming possible.
Wan Streamer is the only end-to-end model that outputs synchronized audio and video, and it does so under 0.6 seconds. Systems like GPT-4o or Moshi handle speech-to-speech interaction but produce no visual output. Avatar systems do produce video, but their latency excludes the external language model, ASR, and TTS they depend on, so their true user-visible latency is higher than shown.