Alibaba's Qwen 3.8 27B open-weight model launches, tops Hugging Face

Qwen 3.8 27B, an Apache 2 licensed 27B parameter vision-capable LLM, scores 52 on the Artificial Analysis Intelligence Index, matching GPT-5.6 Luna (max). It defaults to 'xhigh' reasoning effort, causing overthinking; Simon Willison recommends raising context to 262K tokens.