＋ POSITIVE65%
Alibaba's Qwen team has unveiled Qwen 3.8, a massive 2.4-trillion-parameter AI model that will soon be available under an open-weight license. This move underscores a commitment to transparency and community-driven AI development. The model is already accessible via a preview on Alibaba's Token Plan, Qoder, and QoderWork platforms, allowing developers and researchers early hands-on experience. By making such a powerful model openly available, Alibaba is fostering innovation and collaboration in the AI ecosystem. The Qwen 3.8 is positioned as one of the most capable models today, trailing only a few competitors, and its open-weight release promises to accelerate breakthroughs across industries.
Source weight: ~2 documents
＝ NEUTRAL25%
Alibaba's Qwen team announced the release of Qwen 3.8, a large-scale AI model with 2.4 trillion parameters. The model is set to become open-weight in the near future, with a preview version now live on Alibaba's Token Plan, Qoder, and QoderWork services. According to the announcement, Qwen 3.8 is among the top-performing models currently available, second only to a leading competitor. The open-weight strategy aims to provide broader access to advanced AI technology. The preview allows users to test the model's capabilities before the full open-weight release.
Source weight: ~2 documents
− NEGATIVE10%
Alibaba's Qwen 3.8 launch raises questions about the practicality of ever-larger AI models. With 2.4 trillion parameters, the model's computational and energy demands are immense, potentially exacerbating environmental concerns and limiting accessibility despite claims of open-weight release. The announcement lacks concrete details on when the full open-weight version will be available or under what license. Furthermore, the model's performance claims—being second only to a competitor—are unverified by independent benchmarks. The preview on Token Plan may offer only restricted access, and the hype surrounding such massive models often overshadows more efficient, lightweight alternatives.
Source weight: ~2 documents
