OpenAI has launched GPT-5.2, reporting significant improvements in its ability to handle business tasks at an expert level compared to its predecessor. The new model excels in various benchmarks, including OpenAI’s own GDPval, where it surpassed human expert performance in a majority of tests.
GPT-5.2 offers enhanced capabilities in tasks like spreadsheet formatting, coding, and complex project handling. Despite a slight price increase, OpenAI claims improved efficiency makes it more cost-effective. The release comes amid competition with Google’s Gemini 3, with experts debating the benchmark methodologies.