DigiNews

Summary

Alibaba Qwen announced Qwen3.8-27B, a 27B-parameter native multimodal model that reportedly outperforms Qwen3.7-Plus in coding and office workflows, with a 262K native context extendable to 1M via YaRN. The release emphasizes open weights and Apache 2.0 licensing, with links to HuggingFace and ModelScope for deployment locally or in agents.