AI agents may spend 1,000x more tokens without better results: the hidden bill

花了1000倍的token，效果可能却没有更好：AI Agent的“隐性账单”长什么样

Researchers used OpenHands to analyze traces from 8 frontier models on 500 swe-bench-verified tasks, finding that agentic coding reached a 154:1 input-output token ratio and that human difficulty labels correlated weakly with token use at Kendall tau 0.32.

Why it matters: All HKR axes pass: strong cost-performance hook, concrete benchmark setup and correlation numbers, and direct resonance with coding-agent economics. It is not a model or platform launch, so it fits the 78–84 quality-recommendation band.