Personal intelligence explodes! Ant released a new model to accept the recruitment, and the assistant in the editorial department of the actual test was online

Zhidongxi Author | Yang Jingli Editor | Li Shuiqing Zhidongzhi reported on October 15 that in the past week, major Silicon Valley AI companies have released new developments in turn, all pointing to letting AI take over real work. On October 12, Meta released the Muse Wonder series, including two models, Muse Wonder and Muse Wonder Full, focusing on personal and professional agent tasks. On October 9, Google Cloud also launched the enterprise general intelligent agent Gemini Agent at the Gemini at Work 2026 conference, which can independently plan tasks, call tools, and complete multi-step work in office documents, emails, and development environments. Yesterday, Google opened its latest flagship model Gemini 4 Argon, focusing on complex real-world tasks in programming, finance, law, network security and science. While overseas manufacturers are accelerating the deployment of personal and enterprise intelligence, Chinese manufacturers have continued to explore industrial scenarios such as office, finance, and medical care. How much work a model can complete and the quality of its delivery are becoming important ways to observe AI capabilities. Focusing on real mission requirements, Ant Bailing released Ling-3.1-flash on September 30. Today, Ling-3.1-flash is officially open source. The total parameter volume of the model is about 560B, each Token activation parameter is about 25B, and the upper limit of the context window is 1M. It is focused on daily office work, long-term software development and complex agent tasks. In a public test by Artificial Analysis, a third-party evaluation agency, Ling-3.1-flash has an intelligence index of 41, an output speed of approximately 210 Tokens per second, and a mixed call price of approximately US$0.19 per million Tokens (approximately RMB 1.27). Ling-3.1-flash maintains a high level of intelligence while taking into account faster output speeds and lower call prices, showing a cost-effective advantage. model release