Expired50% confidenceBenchmarkExact time
The Extended Thinking version topped the overall leaderboard on Artificial Analysis's speech-to-speech quality index with a score of 82.6.
1
Sources
50%
Confidence
Short-term (~14 days)
Relevance
9/15/2026
First Seen
Valid until: 9/29/2026(expired)
Sources
Related Entities
Related Claims
UnverifiedFull multimodal large models that support voice input and output score far lower on agent benchmark datasets than mainstream reasoning models.74% similarUnverifiedThe Extended Thinking version scored 68.6% on the τ-Voice benchmark, 35.1% on Sierra's τ-Voice-banking benchmark, and 97.7% on Big Bench Audio.71% similarUnverified使用语音输入编写Prompt理论上可以获得2-4倍的速度提升68% similarUnverifiedActionable dimensions for evaluating voice AI quality include: whether first-response latency is below 500 milliseconds, whether cross-sentence pauses are handled correctly without premature truncation, and whether recognition accuracy remains stable in noisy environments.68% similarUnverifiedAI-generated voice detection technology identifies synthesized speech by analyzing subtle artifacts, spectral anomalies, and unnatural prosodic patterns in audio.68% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/922953API
curl https://kongchang.com/api/v1/knowledge/claims/922953MCP
get_claim(id=922953)