Expired50% confidenceFactExact time
The gemma:2b model may require a response time of several seconds to over ten seconds for local inference on an ordinary machine.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
7/2/2026
First Seen
This claim has expired. It is kept for the record and no longer describes the current state.
Valid until: 9/30/2026(expired)
Sources
Spring AI + Ollama: Run Large Language Models Locally at Zero Cost
bilibili枫叶边城6/3/2026
Related Claims
VerifiedA single LLM inference call can easily take hundreds of milliseconds or even seconds, whereas the execution time of deterministic code is typically at the microsecond level.71% similarUnverified对于BERT-base模型推理会话初始化通常耗时数百毫秒，重复执行会使延迟增加1-2个数量级71% similarUnverifiedThe reasoning process for a single complex code generation request may last 10–30 seconds, during which it continuously occupies high-end GPUs.67% similarUnverifiedRT-2的580亿参数方案推理延迟高达数秒，难以满足机器人实时控制的毫秒级响应需求67% similarVerifiedThe response time tolerance for code completion in AI programming tools is typically within 200–500 milliseconds.65% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/61455API
curl https://kongchang.com/api/v1/knowledge/claims/61455MCP
get_claim(id=61455)