1、2025-06-17Gemini 2.5:Pushing the Frontier withAdvanced Reasoning,Multimodality,LongContext,and Next Generation AgenticCapabilities.Gemini Team,GoogleIn this report,we introduce the Gemini 2.X model family:Gemini 2.5 Pro and Gemini 2.5 Flash,as wellas our earlier Gemini 2.0 Flash and Flash-Lite model

2、s.Gemini 2.5 Pro is our most capable model yet,achieving SoTA performance on frontier coding and reasoning benchmarks.In addition to its incrediblecoding and reasoning skills,Gemini 2.5 Pro is a thinking model that excels at multimodal understandingand it is now able to process up to 3 hours of vide

3、o content.Its unique combination of long context,multimodal and reasoning capabilities can be combined to unlock new agentic workflows.Gemini 2.5Flash provides excellent reasoning abilities at a fraction of the compute and latency requirements andGemini 2.0 Flash and Flash-Lite provide high performa

4、nce at low latency and cost.Taken together,theGemini 2.X model generation spans the full Pareto frontier of model capability vs cost,allowing users toexplore the boundaries of what is possible with complex agentic problem solving.1.IntroductionWe present our latest family of natively multimodal mode

5、ls with advanced reasoning through thinking,long context and tool-use capabilities:Gemini 2.5 Pro and 2.5 Flash and our earlier Gemini 2.0Flash and Gemini 2.0 Flash-Lite models.Together these form a new family of highly-capable modelsrepresenting our next generation of AI models,designed to power a

6、new era of agentic systems.BuildinguponthefoundationoftheGemini1.5series(GeminiTeam,2024),thisGemini2.Xgenerationbrings us closer to the vision of a universal AI assistant(Hassabis,2025).The Gemini 2.X series are all built to be natively multimodal,supporting long context inputs of 1million tokens a