paper-with-me

홈 › Papers

Realtime-Venus: A full-duplex interaction system with asynchronous delegation

2026-09-12 · Ruixiang Zhao, Hualei Wang, Renhe Sun, Enzhi Zhou, Jincenzi Wu, Xujie Song, Kexin Shi, Zihang Liu, Pengcheng Zhu, Jiayi Zhou, Baoyue Zhang, Changhao Zhang, Zitong Wang, Jinhong Wang, Tong Niu, Jingjing Liu, Junan Lin, Haolin He, Hengshuo Chu, Yuhui Chen, Jian Liu, Yuge Huang, Junliang Xing, Yuntao Wang, Weiqiang Wang, Chun Yu, Yuanchun Shi hf

Natural interaction in digital and physical environments requires continuous perception and timely responses. Spoken dialogue relies on acoustic and linguistic cues, while video interaction also requires grounding the conversation in evolving visual context. We present Realtime-Venus, a proactive full-duplex interaction system with two separately trained 9B models: Realtime-Venus-Omni for audio-visual interaction and Realtime-Venus-Audio for spoken interaction. Each model serves as a complete conversational frontend, integrating continuous perception, conversational control, and native speech generation through a shared causal timeline for user inputs, model outputs, and delegation events. A dual-loop runtime coordinates live interaction with background reasoning and tool execution. Foreground interaction continues while Realtime-Venus-Harness executes tasks asynchronously and returns results for integration into the ongoing dialogue. Both models follow a common post-training recipe combining offline understanding, proactive full-duplex trajectories, and delegation workflows. Among the evaluated online models, Realtime-Venus-Omni achieves the highest scores on six of eight video benchmarks, including StreamingBench (70.2%), OVO-Bench (64.7%), and Daily-Omni (81.3%). Across eight audio understanding and spoken question answering benchmarks, Realtime-Venus-Audio leads the compared models on MMAU (78.0%), MMAU-Pro (63.2%), Llama Questions (83.8%), and Speech CMMLU (67.8%), while matching the best VoiceBench AlpacaEval score of 4.81. On Full-Duplex-Bench v1.5, Realtime-Venus-Audio responds to 75% of user interruptions and achieves continuation rates of 97%, 88%, and 86% under backchannels, other-directed speech, and background speech, respectively, exceeding Gemini 3.1 Live and GPT-4o on all three continuation metrics.

📄 PDF Abstract BibTeX arXiv:2609.13814

Code (0)

등록된 구현이 없습니다.

Tasks

Question Answering

Similar Papers 제목 키워드 기반

Liberating LLM Capabilities in Full-Duplex Speech Models

2026-05-04 · Luoyuan Zhang, Bokai Xu, Junbo Cui, Weiyue Sun 외 arxiv

Speech-based large language models are typically constrained to spoken replies, which limits their user-facing outputs to what can be verbalized and suppresses text-native capabilities such as code generation, structured…

Code Generation

StepAudio 3 Realtime Technical Report

2026-09-12 · Bin Lin, Bo Zhao, Boyang Zhang, Boyong Wu 외 hf

Realtime spoken interaction demands deep reasoning, prompt responses, and fluid turn-taking. We present StepAudio 3 Realtime, an audio-language foundation model organized around a continuous listen-converse-think-act loo…

Full-Duplex-Bench-v3: Benchmarking Tool Use for Full-Duplex Voice Agents Under Real-World Disfluency

2026-04-06 · Guan-Ting Lin, Chen Chen, Zhehuai Chen, Hung-yi Lee arxiv

We introduce Full-Duplex-Bench-v3 (FDB-v3), a benchmark for evaluating spoken language models under naturalistic speech conditions and multi-step tool use. Unlike prior work, our dataset consists entirely of real human a…

DuplexCascade: Full-Duplex Speech-to-Speech Dialogue with VAD-Free Cascaded ASR-LLM-TTS Pipeline and Micro-Turn Optimization

2026-03-10 · Jianing Yang, Yusuke Fujita, Yui Sudo arxiv

Spoken dialog systems with cascaded ASR-LLM-TTS modules retain strong LLM intelligence, but VAD segmentation often forces half-duplex turns and brittle control. On the other hand, VAD-free end-to-end model support full-d…

FLEXI: Benchmarking Full-duplex Human-LLM Speech Interaction

2025-09-26 · Yuan Ge, Saihan Chen, Jingqi Xiao, Xiaoqian Liu 외 arxiv

Full-Duplex Speech-to-Speech Large Language Models (LLMs) are foundational to natural human-computer interaction, enabling real-time spoken dialogue systems. However, benchmarking and modeling these models remains a fund…