Unverified50% confidenceBenchmarkExact time
In local conversational scenarios, 15–30 tokens/second is widely regarded as the threshold for a smooth experience.
1
Sources
50%
Confidence
Long-term
Relevance
9/24/2026
First Seen
Sources
Related Entities
Related Claims
VerifiedThe human perceptual tolerance threshold for latency in natural conversation is approximately 300 milliseconds.68% similarVerifiedIn decode speed tests, speeds of approximately 45 tokens/second are achievable within 20,000 tokens, dropping sharply to approximately 20 tokens/second beyond 30,000 tokens.67% similarUnverified用量上限通常以TPM（Tokens Per Minute，每分钟Token数）来衡量66% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/946837API
curl https://kongchang.com/api/v1/knowledge/claims/946837MCP
get_claim(id=946837)