Alibaba Releases Qwen4-Max Foundation Model, Advancing Agent and Coding Capabilities
English text is being prepared.
On November 16, Alibaba released its new-generation foundation model Qwen4-Max, with 2.4T total parameters, native vision capabilities, and support for a million-token context. The company says it achieves higher performance with lower token consumption; in CodeArena, a global blind test for front-end coding, Qwen4 ranks sixth, and it is in the leading tier in agent capability and professional task evaluations such as Agents' Last Exam (ALE) and Vals Index. Qwen4 is optimized for real-world tasks, can call professional tools to handle long-horizon tasks, and is advancing deployment in fields including automotive manufacturing, game development, financial investment research, and scientific research.
Why read this
SelectedAI score 82
The original text provides Qwen4-Max's parameter count, context length, and evaluation rankings, allowing readers to judge the actual standing of its agent and coding capabilities.
Topics
Body
11月16日,阿里巴巴发布全新一代基座大模型Qwen4-Max,总参数2.4T,具备原生视觉能力,支持百万上下文。基于创新的模型架构和训练优化,Qwen4-Max以更少的Token消耗实现了更高的性能,在智能体和编程等关键能力项上持续进化:在权威的全球前端编程盲测CodeArena中,Qwen4位居第六;在Agents’ Last Exam(ALE)、Vals Index等智能体能力及专业任务三方评测中,Qwen4均处于领先梯队。Qwen4面向全领域的真实任务深度优化,可更好地调用专业工具处理长周期任务,推进AI在汽车制造、游戏开发、金融投研及科学研究等专业领域的深入落地。
本文由阿里巴巴提供,量子位获授权转载,观点归原作者所有。
AI News