Google has introduced a new experimental AI model called Gemini 2.0 Flash Thinking Experimental, designed for multimodal understanding, reasoning, and coding. It is part of Google’s AI Studio and aims to tackle complex problems in areas such as programming, math, and physics. This model is built on the recently announced Gemini 2.0 Flash and is similar to OpenAI’s o1 and other reasoning models, which are capable of self-fact-checking to avoid common AI errors.
However, reasoning models like Gemini 2.0 Flash Thinking Experimental often take longer to produce answers, typically requiring several seconds to minutes. The model pauses to consider related prompts and explains its reasoning before summarizing what it believes to be the most accurate response.
The launch of Gemini 2.0 comes amid a surge of reasoning models from various AI labs, including DeepSeek-R1 from DeepSeek and Alibaba's open challenger to o1. Reports indicate that Google has at least 200 researchers dedicated to developing reasoning models, reflecting a significant investment in this technology.
