Google unveils Gemini 2.0 Flash Thinking to compete with OpenAI o1

In its latest push to redefine the AI ​​landscape, Google has announced Gemini 2.0 Flash Thinkinga model of multimodal thinking that can solve complex problems with speed and clarity.

In a post on social network XGoogle CEO Sundar Photos wrote: “Our most thoughtful brand yet :)”

And on production recordsGoogle explains, “Thinking Systems can think deeper in its solutions than the basics Gemini 2.0 Flash model,” once Google’s latest and greatest, was released eight days ago.

The new version supports only 32,000 tokens (approx 50-60 precious leaves) and can generate 8,000 tokens per response. On the Google AI Studio sidebar, the company says it’s good for “multidimensional understanding, reasoning” and “coding.”

Details of the model’s training system, design, certification, and cost have yet to be announced. Currently, it shows a value of zero for each indicator in Google AI Studio.

Approachable and logical thinking

Unlike competitors’ models o1 and o1 mini from OpenAIGemini 2.0 helps users to find his ideas step by step through a drop-down menu, providing clear, transparent information about how the model reaches its end.

By allowing users to see how decisions are made, Gemini 2.0 addresses concerns that have long treated AI as a “black box,” and brings this model—the wording of which is still unclear—to bear with them. some open models offered by competitors.

My simple tests of this model showed that it accurately and quickly (one to three minutes) answered questions that have been popular with other types of AI, such as counting the Rs in the word “Strawberry.” (See image above).

In another test, comparing two decimal numbers (9.9 and 9.11), the model systematically broke the problem into smaller steps, from analyzing whole numbers to comparing decimal places.

These results are supported by an independent third-party analysis from LM Arenawhich named Gemini 2.0 Flash Thinking the first model to succeed in all LLM categories.

Natural support for uploading and analyzing images

In another improvement on the OpenAI o1 family, Gemini 2.0 Flash Thinking is designed to convert images from the jump.

o1 was founded as a text-only format, but has expanded to include image and file analysis. Both types can also return text only, this time.

Gemini 2.0 Flash Thinking is no longer compatible with Google Search, or integration with other Google apps and third-party tools, according to production records.

Gemini 2.0’s Flash Thinking capability expands user experience, enabling it to handle scenarios that involve different types of data.

For example, in one test, the model solved a puzzle that required the analysis of text and visuals, showing its flexibility in combining and reasoning on all types.

Developers can take advantage of this through Google AI Studio and Vertex AI, where the version is available for testing.

As the field of AI becomes increasingly competitive, Gemini 2.0 Flash Thinking may mark the beginning of a new era of problem-solving models. Its ability to deal with different types of data, provide visual recommendations, and act on a large scale as a competitor in the AI ​​thinking market, competing with the OpenAI o1 family and beyond.