← Back to the wire

Qwen 3.8 27B is excellent, but it defaults to wildly overthinking things

SpeculationModelAug 16, 2026

Alibaba's Qwen 3.8 27B is an Apache 2 licensed, vision-capable LLM that defaults to "xhigh" reasoning effort, causing excessive token usage and slow response times on consumer hardware. Testing on a MacBook Pro and NVIDIA DGX Spark via LM Studio showed the model produced strong results in SVG generation and bounding box detection, but the default setting led to a simple circle prompt taking several minutes. The reviewer recommends running the model on low or no reasoning levels initially.

Receipt № 14131 source · awaiting confirmation ◐

Evidence

1source· awaiting independent confirmation

No score is assigned. Sources and their independence are shown in the citation chain below.

Citation chain · 1 source

AlibabaCompanyNVIDIACompanyDGX SparkModelLM StudioCompanyQwen 3.8 27BModelQwen 3.6 27BModelQwen 3.7-PlusModelllama-serverModelMacBook ProModel
Canonical: https://simonwillison.net/2026/Aug/16/qwen-38-27b/