American tech giant Google has released Gemini 2.5, claiming it to be its “most intelligent AI model.” Developed by Google DeepMind, Gemini 2.5 is reported to outperform popular large language models (LLMs) such as OpenAI’s GPT-4.5, Claude 3.7 Sonnet, and Grok 3 on various benchmarks, particularly reasoning.
The news was confirmed by a press release on the company’s official blog stating, “Today we’re introducing Gemini 2.5, our most intelligent AI model. Our first 2.5 release is an experimental version of 2.5 Pro, which is state-of-the-art on a wide range of benchmarks and debuts at #1 on LMArena by a significant margin”.
Gemini 2.5 pro tops the LMArena leaderboard, beating rival models such as GPT and DeepSeek R1. It has also achieved a score of 18.8% on Humanity’s Last Exam benchmark, which tests AI models on a wide range of complex real-world tasks.
See Related: Introducing Gemini 2.0: Google’s Most Capable Model That Can Power AI Agents
Gemini 2.5 pro is said to be built on the best qualities of Gemini such as native multimodality and a long context window. The new model has a context window of 1 million tokens with a 2 million token context window released soon.
Gemini 2.5 pro is currently only available to Gemini Advanced users in Google AI Studio and the Gemini app. It will soon be available on the Vertex AI platform. Google is yet to release the full pricing for Gemini 2.5 Pro.
