蚂蚁集团发布 Ling-3.2-flash 多模态模型，支持百万 token 上下文
来源标题：Ant launches Ling-3.2-flash multimodal model with million-token context
蚂蚁集团于11月5日发布原生多模态模型 Ling-3.2-flash，可接收文本、图像和视频输入。该模型总参数5600亿，每次推理激活250亿参数，支持百万 token 上下文窗口，主打复杂视觉理解、多步智能体任务、3D 理解与交互及医疗多模态应用。模型已通过在线对话界面及兼容 OpenAI、Anthropic 协议的 API 开放，发布后前两周免费。
以下正文同步自 动点科技 TechNode，版权归原站所有，已转换为易读排版。
Ant Group released Ling-3.2-flash on November 5, adding a native multimodal model that accepts text, images, and video. The model has 560 billion total parameters, activates 25 billion per inference step, and supports a one-million-token context window. Ant says the model focuses on complex visual understanding, multistep agent tasks, 3D understanding and interaction, and medical multimodal applications. It is available through an online chat interface and APIs compatible with OpenAI and Anthropic protocols, with free access for the first two weeks after launch. [Ant Ling]
正文由 探How 从来源站点 RSS 同步，内容未经改写；遇到排版缺失或需要图片、视频时请以原文为准。