After Gemma 3 in March, Google has announced the launch of its latest AI model Gemini 2.5 Pro on Tuesday. Compared to the Gemini 2.0 Flash Thinking model, which was released in December 2024, the Gemini 2.5 models feature an enhanced base model with improved post-training.
“For a long time, we’ve explored ways of making AI smarter and more capable of reasoning through techniques like reinforcement learning and chain-of-thought prompting. Building on this, we recently introduced our first thinking model, Gemini 2.0 Flash Thinking,” the company said in a statement.
READ: Google’s AI chatbot Gemini faces accuracy concerns (December 20, 2024)
Google describes the Gemini 2.5 Pro Experimental as its most advanced model for handling complex tasks. It is said to exhibit strong reasoning and coding capabilities. According to Google, the model leads in maths and science benchmarks such as GPQA and AIME 2025. It also scores 18.8% across models without tool use on Humanity’s Last Exam, a dataset designed to capture the human frontier of knowledge and reasoning.
Gemini 2.5 also has a “Show thinking” option wherein it enables users to view the steps taken to provide an output. Google will be integrating these “thinking capabilities” directly into all their new AI models going forward.
Gemini 2.5 Pro is also multimodal in its future and can process text, audio, images, videos and code repositories. The model also currently has a 1 million token context window which is one of the largest context windows available for experimental models on Gemini. The context window will be expanded to 2 million tokens soon, the company has said.
Gemini 2.5 Pro is available now in Google AI Studio and in the Gemini app for Gemini Advanced users, and will be coming to Vertex AI soon. The pricing will also be announced soon, Google said.
On March 30, Google took to X to announce that it’s making Gemini 2.5 Pro (experimental) free for everyone to use, albeit with tighter rate limits for non-subscribers.
Recently, Google has been topping the headlines among developers and enterprises, especially amid fierce competition from competitors like Anthropic, xAI, OpenAI among others.
READ: Apple eager to defend Google during antitrust trial (December 26, 2024)
With its release of Gemini Code Assist, (a product that is powered by Gemini 2.0) Gemma 3, an openweight model and recent updates to Gemini’s Deep Research, a tool that explores complex topics and delivers its findings in a comprehensive, detailed report, and is now available for anyone to try; Google has been constantly trying to position itself as a key player in the AI race.
Additionally, Audio Overviews released in March for Deep Research, where users have the option to listen to an AI-generated, podcast-like discussion of their report, according to Google, has also made its mark on users.
Editor’s note: The above article has been updated to reflect latest information.