Google Releases Gemini 2.5 Pro, Tops Multiple Tests, and Surpasses OpenAI in Reasoning Ability Across the Board

Google has released Gemini 2.5 Pro, the first member of the Gemini 2.5 Think Model family. The model demonstrates outstanding performance in multiple benchmark tests, achieving a score of 1443 in the Large Model Arena and securing a clear first place with a commanding lead of 39 points, comprehensively surpassing several well-known models such as OpenAI o3-mini. In the “Humanity’s Final Exam,” it scores nearly 5% higher than OpenAI o3-mini, representing a 34% improvement. Gemini 2.5 Pro supports a context window of up to 1 million tokens.