Notable model release Alibaba

Alibaba's Qwen Team Launches 27B Parameter Qwen 3.8 Models with Open Weights

Published
Aug 14, 2026 — 17:01 UTC

Alibaba’s Qwen team has released the Qwen 3.8 model, which boasts 27 billion parameters and can process a maximum of 262,000 tokens. This new model outperforms its predecessor, Qwen 3.7 Plus, and is designed for developers focusing on local and agent-based applications. The Qwen 3.8 models are available under the Apache 2.0 license, allowing for broader usage and integration in various projects. This release follows previous developments in the Qwen series, emphasizing Alibaba’s commitment to advancing multimodal AI capabilities. The open weights facilitate easier adoption and experimentation for AI practitioners, enhancing the potential for innovative applications in the field. For more details, see The Decoder.

Turing Wire

By Callan Zhang · Aug 14, 2026 · Editorial standards →

Summarised from the primary source with AI assistance under human editorial oversight. Turing Wire is not a primary source — read the original for the authoritative account.

Source: The Decoder