Google Opens Early Access to Updated Gemini 2.5 Pro AI Model

For developers

Google has launched early access to a new version of its AI model, Gemini 2.5 Pro Preview, which is specifically enhanced for programming tasks. Although the official premiere was planned for the Google I/O conference, due to high interest, the company decided to open access earlier.
Google Opens Early Access to Updated Gemini 2.5 Pro AI Model
© Google

The new model is particularly good at tasks like editing and transforming code, building complex AI agents, and developing interactive web interfaces, said Talsi Doshi, Google’s chief product officer.

The updated Gemini 2.5 Pro has become the leader of the WebDev Arena rating, beating the previous version by 147 points. This rating evaluates how well the model creates beautiful and convenient web applications.

In addition, the model remains strong in working with multimedia data and long texts – it scored 84.8% on the video analysis test.

Developers can already try out the updated model via the Gemini API in Google AI Studio and Vertex AI.

Source: Google

Similar Posts