Google recently announced a major upgrade to its inference model Flash Thinking 2.0, while extending access to the in-depth research feature to all users. This update brings a number of practical features, significantly improving the processing power and user experience of artificial intelligence systems.
According to official introduction, the latest version of Flash Thinking 2.0 has added file upload function and greatly improved processing speed. For Gemini Advanced users, the system is now able to process up to 1 million tagged information in a single context window, meaning that AI can simultaneously analyze and understand larger data sets.
Meanwhile, Google optimized its in-depth research capabilities through Flash Thinking 2.0. The upgraded tool can process information from multiple websites and documents in minutes and show users the process of drawing conclusions through a step-by-step reasoning chain. "By combining in-depth research with this new model, we expect the quality of the report to continue to improve," Google said.
The most significant change in this update is in accessibility – the in-depth research feature is now available to all users and is available to non-paid users several times a month. However, like similar tools from other companies, the system is not perfect. The test results show that in terms of analysis and conclusions, the system still has a certain error rate.
Nevertheless, this upgrade still marks an important progress in Google's AI reasoning capabilities, providing users with more powerful and easier-to-use smart tools.