Qwen3.6-27B Debuts, Delivering Performance Close to Claude Opus 4.5 While Running on a Local PC; Model Available for Anyone to Download

Alibaba’s Qwen team has released Qwen3.6-27B, an open multimodal coding model that can run locally.

The 27B-parameter model is licensed under Apache 2.0, making it suitable for commercial use.

Weights are available on Hugging Face and ModelScope, with FP8 quantization support for easier deployment.

Reported benchmarks suggest it outperforms earlier Qwen and Gemma models, and in some cases rivals Claude Opus 4.5.

A strong open-source local model could reduce deployment costs and broaden access for developers and businesses.