The Google Gemini app introduces access to "thinking" artificial intelligence models.
More effective responses in an instant, even in situations that may be prolonged.
Google has decided to integrate its experimental AI model “reasoning,” capable of explaining its responses to complex questions, into the Gemini application. This update, known as Gemini 2.0 Flash Thinking, is part of a series of releases for the Gemini 2.0 version that the company has recently announced, which also includes its latest flagship model, Gemini 2.0 Pro.
This move comes in the context of Google planning to invest $75 billion in various areas, including the expansion of its family of AI models, representing a significant increase from the $32.3 billion it allocated to capital expenditures in 2023. The company aims to remain competitive against other AI giants like OpenAI, Microsoft, Meta, and Anthropic, backed by Amazon.
The Gemini 2.0 Flash Thinking option will be available starting today on both the desktop and mobile versions of the application, alongside another more interactive variant that allows "interaction with applications like YouTube, Search, and Google Maps." This model was introduced in December 2024 and will compete with other similar AI models, such as OpenAI's o1 and DeepSeek's R1. These models break down problems into more manageable steps, allowing them to "think" about requests before providing solutions, resulting in stronger and more accurate outcomes, although they often require more time to achieve them.
Additionally, an experimental version of Gemini 2.0 Pro is expected to be launched. According to previous leaks, this successor to Gemini 1.5 Pro will offer “better factuality” and “higher performance” in programming and mathematics tasks. It is described as the "most capable model from Google to date" and will be available to Advanced Gemini users and those with access to Vertex AI and AI Studio.
On another note, the latest version of the high-efficiency AI model, Gemini 2.0 Flash, will also be available for developers in AI Studio and Vertex AI after its recent launch in the web and mobile applications of Gemini. Finally, Google has introduced a new low-cost model called 2.0 Flash-Lite, which is said to match the speed and price of 1.5 Flash, surpassing it “in most benchmarks.” This option will make its public preview debut today in Google AI Studio and Vertex AI.