Qwen 3.8 27B Scores 52 on Artificial Analysis

Posts by @itsPaulAi and others state that Qwen 3.8 27B reached a 52 score on artificial analysis benchmarks. The Qwen account described a local 27B model scoring frontier performance and thanked @cline for the mention. @itsPaulAi called it the first time a model runnable locally on a laptop reached parity with leading systems. @ttunguz reported swapping the model into an agent where it performed well. Additional posts from @Alibaba_Qwen and others repeated the local run capability and benchmark result without further details on testing methods or independent verification.