Ant Group's Realtime-Venus: full-duplex interaction with asynchronous model delegation
antgroup · hf · 2026-09-15
Ant Group released Realtime-Venus on Hugging Face, a proactive full-duplex interaction system.
- Architecture: asynchronous delegation separates audio-visual and audio models, integrating continuous perception, conversational control, and native speech generation
- Runtime: a shared causal timeline and dual-loop runtime coordinate the modules
- The system can proactively initiate interaction rather than only responding, targeting real-time voice conversations
More from Multimodal
- GPT-6 Astra + Hyper3D Rodin MCP Turns One Image Into a Full 3D Scene — heyshrutimishra · 2026-09-15
- RTX 5060 runs 12-sec H3 video with 2 ref images in 57 minutes — thatguyjames_uk · 2026-09-15
- 4060 Ti test: LTX 2.5 beats H3 on VRAM and speed, so why does everyone use H3? — ForesterAI · 2026-09-15
- MiniMax H3 hits 2x real-time video denoising with SGLang + VDN-H3 on 8x B200 — MiniMax_AI · 2026-09-15
- Open-source LongCat-Avatar turns one photo plus audio into minutes of lip-synced talking video — Roger_M_Taylor · 2026-09-15
- Odyssey Teases World Models Part 3 Release, Following Interactive Video Push — soleio · 2026-09-15