Google releases Gemini 2.5 Pro "I/O" model; OpenAI plans to cut revenue sharing with Microsoft丨AIGC Daily
-
[OpenAI plans to cut revenue sharing with Microsoft] OpenAI has reportedly told investors that it will reduce the proportion of revenue it shares with Microsoft. The report said that by the end of this decade, the proportion will be at least half. (Sina Finance)
-
[Google releases Gemini 2.5 Pro "I/O" model] On May 6, local time, Google DeepMind's artificial intelligence research team released Gemini 2.5 Pro "I/O", an upgraded version of the multimodal large language model Gemini 2.5 Pro released in March this year. On the AI ranking LMArena, the model topped multiple indicators. The model is currently open to independent developers of Google AI Studio, enterprise users of the Vertex AI cloud platform, and ordinary users of the Gemini application. (Interface News)
-
[Star Era announced the open source of the first AIGC robot large model] On May 7, Star Era announced the open source of the first AIGC robot large model. It is understood that VPP transfers the generalization ability of the video diffusion model to the general robot operation strategy, cleverly solves the problem of diffusion reasoning speed, allows robots to make future predictions and action executions in real time, greatly improves the generalization of robot strategies, and is now fully open source. (IT Home)
-
[Lenovo releases new end-side AI technology] On May 7, at the Lenovo Innovation Technology Conference (Tech World), Lenovo officially released the "Lenovo Tianxi" personal super-intelligence and "Lenovo Lexiang" enterprise super-intelligence. Lenovo demonstrated its latest innovation in end-side computing power, the "Lenovo Reasoning Acceleration Engine". This reasoning engine, jointly developed by Lenovo, Tsinghua University and Wuwen Xinqiong, is a software and hardware collaborative optimization platform designed for efficient AI PC reasoning, which can make the local reasoning ability of an ordinary PC comparable to the o1-mini cloud model released by OpenAI last year. (Securities Times)