Qwen 3.8 27B is out: open weights, best local dense model yet

Alibaba's Qwen has released Qwen3.8-27B, an open-weight dense language model featuring 27 billion parameters with native vision-language capabilities and improved performance across coding, research, and agentic tasks. The model includes flexible thinking control, supports a 262,144 token context length, and is available in FP8-quantized format compatible with major inference frameworks, with a managed API service coming soon.