aifollow.news
蚂蚁百灵发布 Ling-3.2-flash，支持图文视频输入并提供两周免费体验
The Ant Bailing team has released Ling-3.2-flash, a native multimodal model that accepts images, text and video and has a 1M context window. Users can try it on a chat page or call it through an API compatible with OpenAI and Anthropic protocols; access is free for the first two weeks after launch. In the teams published evaluation, the model scored higher than Opus 5 (max) on three listed tests; these are the teams own results.
感谢IT之家网友 软媒用户1773925 的线索投递！
IT之家 11 月 6 日消息，蚂蚁百灵官方团队今日宣布推出原生多模态模型 Ling-3.2-flash，拥有 560B 总参数、25B 激活参数，1M 上下文窗口。
该模型原生支持图像、文本与视频输入，重点强化复杂视觉理解、多模态长程 Agent、3D 理解与交互、医疗多模态四大能力，可通过对话页在线体验，并兼容 OpenAI 及 Anthropic 协议进行 API 调用，而且上线前 2 周限时免费体验。
该模型在发布评估中表现优于 Opus 5（max）：
- OmniDocBench1.5：92.21 vs 90.56
- ZEROBench (Sub)：43.71 vs 39.52
- Vision2Web：85.86 vs 80.20
Name a person, company or topic. We find the sources and gather every new development in your own tab.