Google has today released Gemini 3 Flash, a new 'frontier intelligence built for speed' model, which is seamlessly integrated into the broader Gemini 3 umbrella. This highly efficient model is designed to excel in tasks such as coding, complex analysis, and providing quick answers within interactive applications, all while maintaining high quality. Gemini 3 Flash is a direct successor to Gemini 2.5 Flash, which was launched approximately six months prior.

Built for iterative development, Gemini 3 Flash offers the Pro-grade coding performance found in Gemini 3 Pro but with significantly low latency. This enables the model to reason and solve tasks rapidly in high-frequency workflows. According to Google, Gemini 3 Flash demonstrates a significant performance improvement over its predecessor. On the Humanity’s Last Exam benchmark, designed to test expertise across diverse domains, Gemini 3 Flash scored 33.7% without tool use. This figure surpasses Gemini 2.5 Flash's 11% and closely aligns with the performance of Gemini 3 Pro (37.5%) and OpenAI’s ChatGPT-5.2 (34.5%) in certain measures. Google highlighted that Gemini 3 Flash elevates the performance floor for coding sessions through its strong capabilities in reasoning, tool use, and multimodal functionalities.

Furthermore, Google stated that Gemini 3 Flash has an enhanced ability to understand the intent behind users’ prompts, allowing it to generate more visual answers, including images and tables. As of today, Gemini 3 Flash has been made the default model in the Gemini app and AI Mode in Search, replacing Gemini 2.5 Flash. This integration ensures that developers and everyday users can access the model across various Google platforms. While all users can access Gemini 3 Flash, those on the free mode will have daily usage limits, whereas paid users are granted full access to both Gemini 3 Flash and Gemini 3 Pro.

The introduction of Gemini 3 Flash comes just one month after Google launched its broader Gemini 3 models. These models are characterized as state-of-the-art in reasoning and are built with advanced creativity and problem-solving abilities to grasp depth. The Gemini 3 series was specifically designed to accurately discern the intent behind a user’s request, aiming to fulfill user needs with less prompting. These continuous innovations by Google are unfolding within an increasingly competitive artificial intelligence landscape, marked by the presence of rivals such as OpenAI’s ChatGPT, Zuckerberg’s Meta AI, and China’s DeepSeek. Despite this, the Gemini platform maintains a substantial global reach, serving 2 billion users every month, with the Gemini app alone surpassing 650 million monthly users.