Google Unveils Its Most Advanced AI Model, Gemini 2.5 Pro

Google has just announced the introduction of Gemini 2.5, a significant innovation in artificial intelligence. The first release, Gemini 2.5 Pro, is being touted as a major leap forward in AI reasoning and coding capabilities. This new model comes just three months after the launch of Gemini 2.0, which was designed to enhance complex query handling via AI Mode and Deep Research.

Gemini 2.5 Pro is available as an experimental version on Google AI Studio and is also accessible for Gemini Advanced users on the Gemini chat interface. Koray Kavukcuoğlu, CTO of Google DeepMind, emphasized that Gemini 2.5 is "the most intelligent AI model introduced to date."

The new model has already topped the LMArena leaderboard with a score of 1,443, surpassing competitors such as Grok 3 and GPT-4.5. Notably, Gemini 2.5 Pro has demonstrated exceptional performance in coding, math, and science benchmarks, achieving 84% in GPQA Diamond and 86.7% in AIME 2025.

A core feature of Gemini 2.5 is its innovative reasoning system, which allows the AI to analyze information and draw logical conclusions, resulting in enhanced accuracy and higher performance levels. Google aims to further develop these capabilities to assist with more complex problem-solving, paving the way for more sophisticated, context-aware AI agents.

With the release of the Gemini 2.5 Pro experimental model, developers and enterprises can now start leveraging this new power in Google AI Studio. Google plans to introduce pricing options soon, allowing users to operate 2.5 Pro with higher rate limits for scaled applications.