Google has just unveiled Gemini 2.0 Flash Thinking, an AI model that combines the speed of its Flash technology with the same type of chain-thinking reasoning as OpenAI's o1 model. The new model is available for free to everyone right now on Google's AI Studio platform.
This free version is limited to 32,767 tokens, which is somewhat limited compared to other Gemini models, but what is remarkable is how quickly it tackles complex problems that usually bog down conventional AI models .
Launch demos include puzzles, probability problems, and other examples requiring complex thinking. It's not perfect, some examples from audience members show that it can still make mistakes and miss key aspects of some problem prompts.
The new model has already taken first place in the Lmarena chatbot arena. This is a spectacular result for such a new model, in such a short time. I tested Flash Thinking with a riddle cheekily taken from the OpenAI community's 50 Really Hard Riddles webpage, and it managed to come up with the answer in 9.2 seconds.
While this isn't the kind of test that will prove anything meaningful, it's a pretty decent example of this model's apparent competence.
The other main difference between Google's new reasoning model and OpenAI is the fact that Google AI Studio allows you to adjust security settings to check different answers. So, for example, you can change harassment, hate, dangerous and explicit content and see how these changes affect the model's reasoning abilities.
Those interested in trying the new template for free can sign up on Google's AI Studio platform and select it from the template box in the right sidebar.
News
