
Google has officially released Gemini 2.5, "the most intelligent artificial intelligence (AI) model" to date that pauses to “think” before answering any question.
The model has been launched on Tuesday, March 25, 2025, as Gemini 2.5 Pro in the company’s developer platform, Google AI Studio, and in the Gemini app for subscribers to the company’s $20-a-month AI plan, Gemini Advanced.
Related: Google integrates Gemini's real-time AI video features
Previously, Google has experimented with several AI reasoning models, but Gemini 2.5 represents the company’s most serious attempt yet at besting OpenAI’s “o” series of models.
Gemini 2.5 Pro exceeds its previous AI models on multiple benchmarks. In addition, it is specially designed to surpass at creating appealing web apps and agentic coding apps.
In an evaluation measuring code edition, which is known as Aider Polyglot, the company said that Gemini 2.5 Pro scores 68.6%, exceeding leading AI chatbots, including ChatGPT, Anthropic, and more.
Currently, Gemini 2.5 Pro ships with a one million token context window, which indicates that the AI model can take approximately 750,000 words in a single go.
Notably, Gemini 2.5 Pro pricing is still under wraps; however, it is said to be revealed in the near future.
Related: Google rolls out ‘Canvas’ and ‘Audio Overview’ to Gemini