Qwen 3.6 27B hits 77.2% SWE-bench while staying small enough for single-GPU inference
Alibaba released Qwen 3.6 27B with a 77.2% SWE-bench Verified score — a frontier-competitive number on a model small enough to run on a single H100. The 27B parameter sweet spot has become the most-shipped open-weights size of 2026.
27B has emerged as the most practical open-weights size for the year: too small to need multi-GPU, large enough to host useful coding agents. Qwen 3.6 27B joins Llama 4 Scout, Gemma 4 27B, and Mistral Medium 3.5 in this band.
Apache 2.0-equivalent licensing means the model is shippable in commercial products. Expect to see it embedded in vertical agent products within weeks.