Google has announced a new artificial intelligence (AI) model. It is called Gemini 2.5 Pro. This is the first model in the new Gemini 2.5 series. Rave reviews are beginning to appear.

This "thinking model" is designed to reason and think more deeply. Google says it tries to understand problems better before answering. It analyzes information, finds connections, and uses context. This helps it solve complex tasks more accurately.

Google claims Gemini 2.5 Pro shows big improvements over older models. Its reasoning abilities are much stronger. The new model is also significantly better at coding problems. Google noted a big jump in coding skills from the previous version. It can even create complex code from just a basic idea.

The model understands different kinds of information. It processes text, images, audio, and video inputs together.

A very large memory (context window) can handle 1 million tokens of input. This is like reading thousands of pages at once. It helps the AI understand long documents or conversations. Google plans to make this memory even larger in the future.

Complex, real-world challenges

The model can also give very long and detailed answers. It can output up to 64,000 tokens.

Gemini 2.5 Pro can use search to find current information. It can run code to perform calculations. It can also connect to other software functions.

Tests show the model performs well. It has achieved top scores on many difficult AI benchmarks. It leads in areas like reasoning, coding, and understanding long information.

Its knowledge includes information up to January 2025.

Right now, Gemini 2.5 Pro is available as an experiment. People with a Gemini Advanced plan can use it. Developers can access it through Google AI Studio. It is also available via the Gemini API for building applications. Soon, it will be offered on Google Cloud (Vertex AI) for businesses.

Google sees Gemini 2.5 Pro as a major step. It helps AI become better at handling complex, real-world challenges.