Google introduced a large -scale update of its Gemini artificial intelligence platform, adding an experimental model with the “reasoning” function to the application. This model, called Gemini 2.0 Flash Thinking, explains the course of its reasoning before answering a difficult question. The update has become part of the Gemini 2.0 large release, which also includes the flagship model Gemini 2.0 Pro.
Image source: blog.google
The Gemini 2.0 Flash Thinking model was first represented in December 2024 and was supposed to compete with similar reasoning and models such as O1 from Openai and R1 from Deepseek. According to The Verge, the essence of the system is the ability to divide the tasks into smaller steps before suggesting a solution. This approach, according to the developers, helps to achieve more accurate results, although the final answer can take more time.
New opportunities are already available to users on desktop and mobile devices. Gemini 2.0 Flash Thinking can be selected in the application window through a drop -down list of models. Also, with the update, it became possible to use the functions of artificial intelligence (AI) in other Google applications, such as YouTube, “Search” and Google cards.
In addition to Flash Thinking, the experimental version of Gemini 2.0 Pro was published, which, according to rumors, has higher accuracy and improved characteristics for solving problems related to programming and mathematics, and, according to the company, is the “most powerful model” in the lineup. The PRO version will be available to users of the Advanced Gemini tariff plan, as well as customers who have access to Vertex AI and AI Studio.
The budget model Gemini 2.0 Flash-Lite is also presented, which retains the speed and cost of the predecessor (version 1.5 Flash), but shows the best results in most tests. Flash-Lite is available in public testing mode through AI Studio and Vertex AI.
It is worth noting that against the backdrop of the highest competition and confrontation of companies such as Openai, Microsoft, Meta✴ and Anthropic, Google actively increases investment in the development of artificial intelligence. In 2025, the company plans to spend $ 75 billion for these purposes, which is more than twice the expenses of 2023 ($ 32.3 billion), and the Gemini 2.0 renewal is part of these plans.