Google DeepMind announced Gemini 3 Flash, a new AI model blending frontier intelligence with Flash-level speed. Designed for rapid coding, complex analysis, and real-time interactions, it outperforms Gemini 2.5 Pro in benchmarks while cutting token costs by 30%. Developers and consumers can now access it via the Gemini app, AI Mode in Search, and tools like Google AI Studio and Vertex AI.

Gemini 3 Flash achieves 90.4% on GPQA Diamond and 81.2% on MMMU Pro, rivaling larger models. It processes queries 3x faster than predecessors at a fraction of the cost—$0.50/1M input tokens and $3/1M output tokens. Its efficiency stems from dynamic reasoning modulation, adapting compute power to task complexity. On SWE-bench Verified, it scores 78%, surpassing Gemini 3 Pro in coding agility, enabling applications like in-game assistants and data-driven experiments.

The model powers Google’s AI Mode in Search, delivering multimodal reasoning for tasks like trip planning or educational Q&A. By integrating real-time web data and visual analysis, it transforms unstructured voice inputs into functional apps. Enterprises via Gemini Enterprise and Vertex AI already report transformative results, with companies like JetBrains and Figma adopting it for scalable workflows.

Gemini 3 Flash democratizes advanced AI, offering Pro-tier capabilities at no cost through the Gemini app. Its rollout prioritizes speed without sacrificing depth, positioning it as a versatile tool for developers, researchers, and everyday users navigating complex digital landscapes.