Google has started the rollout of Gemini 2.5, which the tech company claims to be the “most advanced AI model” developed by its DeepMind research lab for the Gemini AI chatbot.
Gemini was launched in December 2023 as Google’s competitor to ChatGPT, which was released by Microsoft-backed OpenAI in November 2022. It has been operating on the Gemini 2.0 model that debuted in December 2024.
The initial release of 2.5 is an experimental version of Gemini 2.5 Pro, which Google states features enhanced cognitive abilities.
“Gemini 2.5 models are designed for reasoning, allowing them to process their thoughts before providing responses, leading to better performance and greater accuracy,” said Koray Kavukcuoglu, CTO of Google DeepMind, in a blog post.
Gemini 2.5 Pro is accessible through Google AI Studio and the Gemini app for advanced users. Google has indicated that this model will soon be available on its machine learning platform, Vertex AI.
“We will also announce pricing in the upcoming weeks, allowing users to access 2.5 Pro with increased rate limits for larger-scale production applications,” Kavukcuoglu added.
According to Google, the new model demonstrates impressive reasoning skills and excels in coding, mathematics, and scientific benchmarks.
“It can analyze extensive datasets and tackle intricate problems from various information sources, including text, audio, images, video, and even complete code repositories,” Kavukcuoglu explained.
Previously, Gemini 2.5 Pro was exclusively available to Gemini Advanced subscribers as part of the Ksh.3,700-a-month Google One AI Premium plan.
Google has been incorporating its AI advancements into its range of products, including Search, Android, and Workspace.
For example, the Gemini 2.0 Flash model has been driving the AI Overviews feature in Search.
This initiative is part of the tech giant’s strategy to compete with companies like OpenAI, which offers a Ksh.26,000-a-month ChatGPT subscription for advanced research purposes.
In December, the company introduced its text-to-video model, Sora, for ChatGPT Plus and Pro users. This feature enables users to create videos with a resolution of up to 1080p, lasting 20 seconds, and available in different aspect ratios.
Among the competitors challenging Google in the AI landscape is Elon Musk’s xAI, which has integrated its chatbot Grok into the X social media platform.
In January, China’s DeepSeek made waves in the industry by unveiling a model that it claims was trained using less sophisticated, more affordable chips compared to those of its Western competitors.
