inclusionAI Releases Duo v1, a Full-Duplex Audiovisual World Model With 93.3% Pass Rate at Sub-650ms Latency

inclusionAI released Duo v1 on Nov 13, pairing an autoregressive 'intent' multimodal LLM that decides when to speak with a causal audiovisual diffusion transformer that generates synchronized video and audio in 0.71-second chunks. On 30 real-time duplex interactions the system cleared 93.3% of 180 pass criteria, hitting 100% on language adherence and speech intelligibility and 90% on lip-sync, with 330-650ms time-to-first-output on 8 GPUs.