Local models handle half of daily tasks and respond faster than cloud models

本地模型处理半数日常任务，响应速度优于云端

A five-week experiment tested about 1,400 daily work tasks, where local 35B models such as Qwen 3.6 35B handled about 50% and averaged 2.8-second responses, 2.1 times faster than Claude Opus 4.5, while the cloud model still led complex reasoning by about 20%.

Why it matters: HKR-H/K/R all pass: Tom Tunguz’s experiment reports ~1,400 tasks, ~50% success, 2.8s latency, and a speed comparison to Claude Opus 4.5. Strong practitioner signal, but not a model launch or platform-level update.