In plain words: A new family of Gemini models spans cheap, fast versions to a top model that thinks before answering and can read up to three hours of video. The top model leads coding and reasoning tests, while smaller ones cut compute and latency sharply.
Abstract · Gemini 2.5: Pushing the Frontier with Advanced Reasoning, Multimodality, Long Context, and Next Generation Agentic Capabilities
In this report, we introduce the Gemini 2.X model family: Gemini 2.5 Pro and Gemini 2.5 Flash, as well as our earlier Gemini 2.0 Flash and Flash-Lite models. Gemini 2.5 Pro is our most capable model yet, achieving SoTA performance on frontier coding and reasoning benchmarks. In addition to its incredible coding and reasoning skills, Gemini 2.5 Pro is a thinking model that excels at multimodal understanding and it is now able to process up to 3 hours of video content. Its unique combination of long context, multimodal and reasoning capabilities can be combined to unlock new agentic workflows. Gemini 2.5 Flash provides excellent reasoning abilities at a fraction of the compute and latency requirements and Gemini 2.0 Flash and Flash-Lite provide high performance at low latency and cost. Taken together, the Gemini 2.X model generation spans the full Pareto frontier of model capability vs cost, allowing users to explore the boundaries of what is possible with complex agentic problem solving.
Gheorghe Comanici, Eric Bieber, Mike Schaekermann, Ice Pasupat, Noveen Sachdeva, Inderjit Dhillon, Marcel Blistein, Ori Ram, Dan Zhang, Evan Rosen, Luke Marris, Sam Petulla, et al.
arXiv:2507.06261 · cs.CL, cs.AI · submitted Jul 7, 2025 · updated Dec 19, 2025
abstract · pdf · html · 72 pages, 17 figures
> The development of Gemini is a large-scale collaborative effort involving over 3000 individuals across Google, including researchers, engineers, and operations staff. These individuals contributed their hard work and expertise across diverse areas, from foundational research and the development of model architecture, data, training, and infrastructure, through to evaluation and ensuring safety and security. We gratefully acknowledge the dedication and hard work of each contributor in making Gemini a reality.