Full-Duplex-Bench-v3: Benchmarking Tool Use for Full-Duplex Voice Agents Under Real-World Disfluency
Guan-Ting Lin, Chen Chen, Zhehuai Chen, Hung-yi Lee
Recommendation Score
Research context
Topics
Tool Use, AI Agents
Paper type
Benchmark
Best for
Useful for both
arXiv categories
Why It Matters
Voice agents often fail when users stutter, pause, or interrupt, leading to broken API calls and frustrated users. This benchmark uses real human speech to reveal exactly how top models handle these messy realities. It allows developers to test if their voice systems can actually execute tasks reliably in natural conversation.
Abstract
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 audio annotated for five disfluency categories, paired with scenarios requiring chained API calls across four task domains. We evaluate six model configurations -- GPT-Realtime, Gemini Live 2.5, Gemini Live 3.1, Grok, Ultravox v0.7, and a traditional Cascaded pipeline (Whisper$\rightarrow$GPT-4o$\rightarrow$TTS) -- across accuracy, latency, and turn-taking dimensions. GPT-Realtime leads on Pass@1 (0.600) and interruption avoidance (13.5\%); Gemini Live 3.1 achieves the fastest latency (4.25~s) but the lowest turn-take rate (78.0\%); and the Cascaded baseline, despite a perfect turn-take rate, incurs the highest latency (10.12~s). Across all systems, self-correction handling and multi-step reasoning under hard scenarios remain the most consistent failure modes.