Google has introduced Gemini 2.5, a cutting-edge family of AI reasoning models designed to push the boundaries of artificial intelligence capabilities. The current release includes the Gemini 2.5 Pro. It’s heavy hitting with an amazing 1 million token context window—and AI able to process 750,000 words in one prompt—that’s enough to read the entire Lord of The Rings book series in one go! Gemini 2.5 Pro hits on Tuesday, join us! You can use it either through Google AI Studio or by signing up for the company’s $20-a-month AI plan, Gemini Advanced.
Whereas the default Gemini 2.5 Pro model is designed to be strong at producing visually rich web applications and agentic coding applications, Google claims that this new model surpasses its prior frontier AI models. It beats many of its top competitors on a variety of other science-based benchmarks. Specifically, Gemini 2.5 Pro achieves a remarkable 68.6% on the Aider Polyglot benchmark, a test of code editing functionality. This achievement puts it on the heels of leading, humanlike performance by state-of-the-art AI models from OpenAI, Anthropic, and Chinese AI laboratory DeepSeek.
We hope to see you virtually! Additionally, the model performs impressively on Humanity’s Last Exam, a multimodal benchmark with thousands of crowdsourced questions, where it scores 18.8%. This achievement further cements its position above prominent competitors in the AI landscape.
Google has already indicated that reasoning will be a built-in, distinguishing feature of all its future models. The first model to demonstrate this is Gemini 2.5. The company recently teased some extremely cool plans for their next upgrade to Gemini 2.5! We’re still finalizing the details, but this upgrade will include a 2m token context window, providing even more processing power and functionality.