Qwen3.8-27B release: New open-weight model for local deployment

Qwen3.8-27B weights have been released. This is a 27B multimodal dense model from Alibaba, runs on 17GB RAM/VRAM, outperforms Qwen3.7-Plus, and supports 262K context. Highly suitable for 32-64GB setups. Qwen3.8-Max (2.4T, 95B active) is far beyond scope but open weights expected next week.