Verified65% confidenceFactExact time
The floating-point operation count of prefill grows approximately quadratically with input length, and is therefore the main source of TTFT (time to first token).
3
Sources
65%
Confidence
Long-term
Relevance
10/14/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedLonger context means higher time-to-first-token (TTFT), and token count is directly linearly correlated with inference costs.75% similarUnverifiedtokens/s指标可细分为首token延迟（TTFT）和生成速率（Generation Throughput）两个子维度75% similarVerifiedInference latency is typically decomposed into two core metrics: time to first token (TTFT) and subsequent token generation throughput.72% similarUnverifiedThe computational cost of Prefill scales approximately linearly to quadratically with input length, making it one of the main sources of inference latency.72% similarUnverifiedThe prefill stage refers to the process in which the model processes the input prompt all at once; the computation is proportional to input length, and in long-document scenarios the time to first token (TTFT) is significantly prolonged.71% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1032220API
curl https://kongchang.com/api/v1/knowledge/claims/1032220MCP
get_claim(id=1032220)