Google has officially launched Gemini 3 Flash, a new AI model designed to deliver faster performance at a lower cost. Built on the Gemini 3 architecture, the company says the Flash variant is up to three times faster than Gemini 2.5 Flash and outperforms all previous Flash models across Google’s internal benchmarks.
In third-party-style evaluations, Gemini 3 Flash has shown surprisingly strong results for a speed-focused model. According to Google, it performs on par with both Gemini 3 Pro and OpenAI’s GPT-5.2 in several benchmark tests. In the multimodal MMMU-Pro benchmark, Gemini 3 Flash reportedly took the top spot with a score of 81.2 percent, highlighting its strengths in handling complex, mixed-input tasks.
Google positions Gemini 3 Flash as a model optimized for high-frequency, fast-turnaround workloads. It is particularly aimed at use cases such as video analysis, large-scale data extraction, and visual queries. The model is also said to be better at understanding user intent and generating more visual responses, including structured outputs like images, charts, and tables.
With this launch, Google is making Gemini 3 Flash the default model in both the Gemini app and Google’s AI-powered search mode. The move signals Google’s push to balance performance and efficiency, bringing advanced AI capabilities to everyday interactions without the overhead of heavier flagship models.
