Google launches Gemini 3 Flash, makes it the default model in the Gemini app (techcrunch.com)

Google has launched its new Gemini 3 Flash model, positioning it as the default in the Gemini app and AI search mode. This model boasts significant performance improvements over its predecessor, Gemini 2.5 Flash, scoring 33.7% on the Humanity’s Last Exam benchmark—near the level of leading models like Gemini 3 Pro and GPT-5.2, while excelling in multimodality with an impressive 81.2% on the MMMU-Pro benchmark. This release marks Google's strategic move to enhance its competitive standing against OpenAI, amid increasing demand for capable and efficient AI models. Gemini 3 Flash is designed for a variety of tasks, including identifying and analyzing multimodal content, making it adept at handling user-uploaded videos, sketches, and audio recordings. It enables users to generate rich visual answers and develop app prototypes through the Gemini app. With an average of 30% fewer tokens used for thinking tasks compared to the previous model, the pricing structure remains competitive, allowing companies to utilize the model for bulk tasks efficiently. As Google accelerates its release of new models in response to the intensifying rivalry with OpenAI, the launch of Gemini 3 Flash signals a notable enhancement in capabilities that could reshape user experiences across various sectors.