Google's Gemini 2.5 Pro has emerged as a leader in coding benchmarks and MENSA IQ tests, outperforming competitors like Claude. It features a one million token context window, expandable to two million, enabling it to manage large codebases effectively. Gemini achieved notable scores on reasoning benchmarks, including an IQ score of 115, categorizing it among high-performing AIs. It scored 86.7% on the AIME 2025 math test and 84.0% on the GPQA science assessment, showcasing its advanced problem-solving capabilities. The model is now free for users with certain rate limits, offering cost-effective options for developers. Its advanced capabilities allow it to create complex applications and assist in coding tasks with impressive accuracy. This release represents Google's strategy to establish Gemini as a versatile AI tool for developers.

Source 🔗