On February 20, Google unveiled Gemini 3.1 Pro, surpassing competitors like Gemini 3 Pro and GPT-5.2 in various internal benchmarks. It showcased an impressive demo by generating a WebOS-like interface for Windows 11 in-browser, complete with interactive elements. Blogger Chetaslua noted that the new model enhances complexity and interaction compared to Gemini 3.0 Pro. The Gemini 3.1 Pro led in 12 benchmarks testing logical thinking and multimodality but performed lower in real engineering tasks, such as SWE-Bench Pro, indicating challenges with large code bases. A key highlight is its ability to manage complex reasoning for technical and creative tasks, making it suitable for sophisticated projects. In demonstrations, it created interactive city simulators, a 3D bird flight simulator, and a dynamic Solar System model. Available to Google AI Pro and Ultra subscribers through apps and APIs, free users can make limited requests. Enjoy innovative tech news and updates by subscribing to our channels!
Source link
Share
Read more