Google introduces Gemini 2.0 Flash Thinking to rival OpenAI o1

In its latest effort to redefine the AI ​​landscape, Google has announced Gemini Lightning Thinking 2.0a multimodal reasoning model capable of addressing complex problems quickly and transparently.

in a post on social networkGoogle CEO Sundar Pichai wrote that it was: “Our most thoughtful model yet :)”

and in the developer documentationexplains Google, “Thinking Mode is capable of having stronger reasoning capabilities in its responses than the base model Gemini 2.0 Flash,” which was previously Google’s latest and greatest, released just eight days ago.

The new model supports only 32,000 input tokens (approximately 50-60 pages of text) and can produce 8000 tokens per output response. In a side panel about Google AI Studio, the company claims it’s best for “multi-modal understanding, reasoning” and “coding.”

Full details of the model’s training process, architecture, licensing and costs have not yet been released. Right now, it shows zero cost per token in Google AI Studio.

Accessible and more transparent reasoning

Unlike OpenAI’s competing o1 and o1 mini reasoning models, Gemini 2.0 allows users to access their reasoning step-by-step through a drop-down menu, offering a clearer and more transparent view of how the model arrives at their conclusions.

By allowing users to see how decisions are made, Gemini 2.0 addresses long-standing concerns about AI operating as a “black box” and brings this model (licensing terms are still unclear) on par with others. open source models presented by the competition.

My first simple tests of the model showed that it answered correctly and quickly (in one to three seconds) some questions that have been notoriously difficult for other AI models, such as counting the number of R’s in the word “Strawberry.” (See screenshot above).

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

These results are supported by independent third-party analysis of L.M. Sandwhich named Gemini 2.0 Flash Thinking as the best performing model in all LLM categories.

Native support for image loading and analysis.

In a further improvement over the rival OpenAI o1 family, Gemini 2.0 Flash Thinking is designed to process images right out of the box.

o1 was launched as a text-only model, but has since expanded to include file and image upload analysis. Both models can also return only text, at this time.

Gemini 2.0 Flash Thinking also does not currently support grounding with Google Search or integration with other Google apps and external third-party tools, according to the developer documentation.

Gemini 2.0 Flash Thinking’s multimodal capability expands its potential use cases, allowing you to address scenarios that combine different types of data.

For example, in one test, the model solved a puzzle that required analyzing textual and visual elements, demonstrating its versatility in integrating and reasoning between formats.

Developers can take advantage of these features through Google AI Studio and Vertex AI, where the model is available for experimentation.

As the AI ​​landscape becomes increasingly competitive, Gemini 2.0 Flash Thinking could usher in a new era for problem-solving models. Its ability to handle diverse types of data, deliver visible reasoning, and operate at scale positions it as a serious contender in the reasoning AI market, rivaling OpenAI’s o1 family and beyond.