According to the announcement, this is an upgraded trial version, following the 2.5 Pro version launched about a month ago. The model is now available on Google's artificial intelligence (AI) development platforms such as AI Studio, Vertex AI and the Gemini app. The official version is expected to be widely released in the next few weeks.
Gemini 2.5 Pro continues to outperform in difficult programming tests, Google wrote in a blog post. The model is also highly effective in complex assessments of Mathematics, science, knowledge and logical thinking.
Google said it has adjusted based on feedback from the previous version, making the response model clearer, better presented, and more creative.
Continuously improving Gemini 2.5 Pro shows Google's ambition in the AI race with rivals such as OpenAI and Microsoft. As the model's programming capabilities, thinking and creativity are improved, Google expects Gemini to become a powerful support tool for both developers and general users in many fields.
