Unverified50% confidenceFactExact time
大多数模型在 NPU 上的实际算力利用率仅为理论峰值的 30%-60%
1
Sources
50%
Confidence
Long-term
Relevance
8/18/2026
First Seen
Sources
Related Claims
Unverified业界顶级集群的MFU（模型算力利用率）通常在35%-60%之间71% similarUnverifiedIn a complete AI application system, GPU inference typically accounts for only 30%–40% of total compute, with the remaining 60%–70% of the compute load borne by the CPU.71% similarUnverifiedMid-tier models typically achieve more than 80% of flagship model scores on benchmarks such as MMLU and HumanEval, while requiring less than 30% of the inference cost.71% similarUnverified一个61分的模型和一个59分的模型在多数实际应用中的差异可能远小于分数暗示的程度71% similarUnverified相较于用GPU执行AI推理，NPU的功耗通常只有前者的1/5到1/1070% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/770835API
curl https://kongchang.com/api/v1/knowledge/claims/770835MCP
get_claim(id=770835)