Unverified50% confidenceBenchmarkExact time
ElevenLabs的Turbo模型支持首句合成延迟约300ms
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
7/23/2026
First Seen
Valid until: 10/21/2026
Sources
语音AI Agent架构选型：Pipecat Flows vs Vapi Squad深度对比
redditr/aiagents7/12/2026
Related Claims
UnverifiedDeepgram Nova-3的流式识别延迟可低至100ms以内，ElevenLabs的Turbo模型支持首句合成延迟约300ms80% similarUnverified大模型调用是典型的I/O密集型操作，一次大模型API调用的延迟通常在1-30秒之间68% similarUnverified小模型推理延迟约500ms，而向量匹配延迟仅30-100ms67% similarUnverified当前主流大模型在处理复杂代码任务时，首token延迟（TTFT）往往在2-8秒之间66% similarUnverified通过中转访问海外模型的延迟通常为300-800ms66% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/599681API
curl https://kongchang.com/api/v1/knowledge/claims/599681MCP
get_claim(id=599681)