Google has introduced its latest Gemini 3 Flash model, a successor to the Gemini 3, as part of its strategy to enhance AI capabilities. The tech company has designated this new model as the default option in the Gemini app and AI mode for search functions, aiming to compete with industry players like OpenAI. Building on the success of the previous Gemini 2.5 Flash model, the Gemini 3 Flash showcases significant advancements. In performance tests, it not only surpasses its predecessor but also competes favorably with cutting-edge models such as Gemini 3 Pro and GPT 5.2 in certain aspects.
For instance, on the Humanity’s Last Exam benchmark, which assesses proficiency across diverse domains without external tools, the Gemini 3 Flash achieved a score of 33.7%. In comparison, Gemini 3 Pro scored 37.5%, Gemini 2.5 Flash scored 11%, and GPT-5.2 scored 34.5%. Moreover, on the multimodality and reasoning benchmark MMMU-Pro, the new model excelled with a score of 81.2%, outperforming all competitors.
Google plans to implement the Gemini 3 Flash as the default model in the Gemini app globally, replacing the Gemini 2.5 Flash. Users will still have the option to select the Pro model for math and coding queries. The company highlights the model’s proficiency in recognizing multimodal content to provide tailored responses. Users can leverage features like uploading videos for tips, sketching for identification, or submitting audio recordings for analysis or quiz generation. Additionally, Google emphasizes the model’s improved query interpretation capabilities, enabling more visual responses with elements like images and tables.
Source: TechCrunch
Leave a Reply