Google is bringing The experimental model of artificial intelligence “reasoning”, which can explain how it answers the intricate questions of the Gemini application. The Flash Gemini 2.0 thinking update is part of a lot Implementation of AI Gemini 2.0 announced by Google todayIncluding the latest flagship model Gemini 2.0 Pro.
This happens when the search giant expects to invest $ 75 billion in expenses, such as cultivation this year a monotonously named AI family. This is a significant jump from $ 32.3 billion on investment expenses, which it spent in 2023, and Google was now racing to keep up with AI competitions, such as OpenAI, Microsoft, Meta and supported by Amazon Antropic.
Flash Gemini 2.0 thinking will be available in the developed options of the model on desktop computers and mobile application from today, along with another, more agency version of the model that can “interact with applications such as YouTube, search and Google maps”, according to Google. It was introduced in December 2024 and is expected to compete with other so -called AI models, such as O1 O1 and Deepseek R1.
These models work by spreading problems into smaller, managing steps, allowing them to “think” about hints before offering the solution. The intended result is to achieve stronger and more correct results, but it is often a longer cost to achieve them.
Google also releases the experimental version of Gemini 2.0 Pro. According to preview leaks Reported by TechCrunchThe successor of Gemini 1.5 Pro should provide “better factual” and “stronger performance” in coding and mathematics tasks. Gemini 2.0 Pro is described as “the most talented model” Google and will be available to advanced users of the Gemini application and people with access to Vertex AI and AI Studio.
Gemini 2.0 Flash-The ONE version of the high performance of the AI Workhorse Google-is also generally available to programmers in AI Studio and Vertex AI after introducing the Gemini internet and mobile applications last week.
Finally, Google completes these updates by introducing a novel inexpensive model called 2.0 Flash-Lite, which according to the company fits up to 1.5 flash and prices, while exceeding it “at most reference points”. Gemini 2.0 Flash-Lite begins today in a public preview in AI Studio Google and Vertex AI.