Skip to content
AI IntelligenceAug 22, 2026AI Intelligence
Article

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

Alibaba's Qwen3.8-27B, an Apache 2.0-licensed 27B vision-capable model, shows large self-reported benchmark gains over Qwen 3.6 27B and the closed-weight Qwen 3.7-Plus. In practice it defaults to excessive reasoning, adding latency and cost. Developers deploying it locally should account for that overthinking behavior.

Frontier EditorialSource: Simon Willison
01

Source Brief

Qwen 3.8 27B is excellent, but it defaults to wildly overthinking things: Alibaba's Qwen3.8-27B, an Apache 2.0-licensed 27B vision-capable model, shows large self-reported benchmark gains over Qwen 3.6 27B and the closed-weight Qwen 3.7-Plus. In practice it defaults to excessive reasoning, adding latency and cost. Developers deploying it locally should account for that overthinking behavior.