Alibaba has unveiled its next-generation foundation model, Qwen4-Max, featuring 2.4 trillion total parameters, native vision capabilities, and support for million-token context. Leveraging an innovative architecture and training optimizations, the model delivers stronger performance with lower token consumption and excels in key capabilities such as coding and agent-based tasks: it ranks sixth in the CodeArena global front-end coding blind test and is among the top tier in third-party evaluations of agents and professional tasks such as ALE and Vals Index. The model has been deeply optimized for real-world tasks across domains, enabling it to better invoke professional tools to handle long-cycle tasks and driving AI deployment in professional fields such as automotive manufacturing, game development, financial investment research, and scientific research. Qwen4-Max is now live and available, with usage costs significantly lower than those of the previous generation, and the small and medium-sized models in the Qwen4 series will be open-sourced and released successively.
