Google has introduced Gemini 2.0 Flash Thinking, a new experimental AI model that “unlocks deeper reasoning capabilities and demonstrates its thoughts.”
Just when you thought it was over… we’re introducing Gemini 2.0 Flash Thinking, a new experimental model that unlocks stronger reasoning capabilities and shows its thoughts.
The model plans (with thoughts visible), can solve complex problems with Flash speeds, and more ?
— Logan Kilpatrick (@OfficialLoganK) December 19, 2024
The neural network is described as “optimal for multimodal understanding, reasoning, and programming with the ability to think about the most complex problems.” It is available in Google AI Studio and Gemini API.
Jeff Dean, Chief Scientist at Google DeepMind, cited the solution of a physics problem as an example and noted that Gemini 2.0 Flash Thinking is “trained to use thoughts to strengthen its reasoning.”
Want to see Gemini 2.0 Flash Thinking in action? Check out this demo where the model solves a physics problem and explains its reasoning. pic.twitter.com/Nl0hYj7ZFS
— Jeff Dean (@JeffDean) December 19, 2024
“We see promising results when we increase computation time in inference mode,” he emphasized.
The new model is developed based on the recently introduced Gemini 2.0 Flash. The neural network is similar to o1 from OpenAI and other so-called reasoning models. They take more time to respond, cross-checking information.
Earlier, Chinese company Alibaba introduced a “thinking-oriented” AI model QwQ-32B-Preview.
In December, OpenAI released the full version of o1 and a professional mode for it.