Alibaba has announced the preview release of QwQ-Max-Preview, the latest addition to the Qwen model series, built on Qwen2.5-Max. This version showcases enhanced capabilities in deep analysis, mathematics, programming, and general tasks, with a focus on agent-related workflows. The company is offering users the opportunity to explore its potential while the team continues to refine its features.
<think>…</think> QwQ-Max-Preview
— Qwen (@Alibaba_Qwen) February 24, 2025
Qwen Chat: https://t.co/FBpr7zfQY6
Blog: https://t.co/Zta7ezuIiC
🤔 Today we release "Thinking (QwQ)" in Qwen Chat, backed by our QwQ-Max-Preview, which is a reasoning model based on Qwen2.5-Max. This model is still for preview. It is highly… pic.twitter.com/eIRP8XGLv2
Alibaba has confirmed plans to open source QwQ-Max and Qwen2.5-Max under the Apache 2.0 license, emphasizing its commitment to fostering collaboration and accessibility in AI development.
Additionally, Alibaba plans to release smaller open versions of its models, such as QwQ-32B, optimized for local deployment. These lightweight models are designed for privacy-sensitive applications and low-latency environments, giving developers and researchers the flexibility to create custom solutions without significant computational costs.
