Unverified50% confidenceBenchmarkExact time
In actual testing, the X-Envoy-Upstream-Service-Time for pure business logic was 33–34 milliseconds, while a coin-flip request that went through full LLM inference took approximately 820–866 milliseconds.
1
Sources
50%
Confidence
Short-term (~14 days)
Relevance
10/11/2026
First Seen
Valid until: 10/25/2026
Sources
Related Entities
Related Claims
UnverifiedAplomb 1 对短问题的模型端计算时间约 15 毫秒，端到端延迟约 200 毫秒70% similarVerifiedOn a T4 GPU, Laya takes 33 milliseconds per single question, and batch processing can compress that to 7 milliseconds69% similarUnverifiedIn this test, when the 7900XTX is run with a 256K context, a single generation takes up to 743 seconds.67% similarUnverifiedThis ComfyUI workflow is adapted from the earlier 'dual-ten sampling' workflow, and raises the 15-second limit to 30 seconds, 40 seconds, or even longer through a segmented continuation approach.67% similarUnverifiedA 3-billion-parameter Llama takes 1 to 3 seconds per call on CPU, while a 70B model takes 40 seconds per call.67% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/995837API
curl https://kongchang.com/api/v1/knowledge/claims/995837MCP
get_claim(id=995837)