Two years after launching its generative AI assistant Bard — later rebranded as Gemini — Google has now introduced what it calls its most advanced model yet: Gemini 2.5 Pro
According to Deccan Herald, this upgraded version boasts significant improvements, especially in reasoning ability. Unlike its predecessors, Gemini 2.5 Pro is designed to "think through" queries before responding, a feature that reportedly results in better accuracy and performance.
The model currently tops the LMArena leaderboard — a global ranking system for Large Language Models (LLMs) and AI chatbots. Developed by researchers from the UC Berkeley SkyLab and the Large Model Systems Organisation (LMSYS), the LMArena platform uses real-time human feedback to assess models using the Bradley-Terry statistical model, Deccan Herald notes.
Google says Gemini 2.5 Pro can do more than just answer questions; it can write and test complex code, solve advanced mathematical problems, and even generate a complete video game from a single-line prompt. The model also supports multimodal inputs, meaning it can understand and respond to text, images, and videos.
Already integrated across the Google Workspace suite — including Gmail, Docs and other tools — the Gemini platform has been aimed at boosting productivity for everyday users. With Gemini 2.5 Pro, the capabilities now extend further into the developer and research space.
It is currently available on Google AI Studio and through the Gemini app for Gemini Advanced users. Google says it will also be added to Vertex AI, the company’s cloud-based machine learning platform, in the near future.
