According to @TheRundownAI, OpenAI has released GPT-5.2 Thinking, marking its first AI model to reach human-expert performance on GDPval, a standardized evaluation spanning 44 professional occupations such as presentations and spreadsheet analysis. The company emphasizes significant advancements in long-context reasoning, coding, scientific workflows, and technical writing. GPT-5.2 Thinking is now accessible via the API and Codex, while GPT-5.1 will remain available for three months, enabling businesses immediate access to enhanced productivity tools and integration pathways (source: @TheRundownAI).

According to OpenAI, GPT-5.2 Thinking is the first AI model to reach human expert-level performance on GDPval, an evaluation that measures well-specified knowledge work tasks across 44 occupations. These tasks cover practical business functions such as making presentations, creating spreadsheets, and producing other professional artifacts, highlighting GPT-5.2's practical capabilities for enterprise automation and productivity (source: OpenAI @OpenAI, Dec 11, 2025). This breakthrough demonstrates significant potential for AI-driven automation in knowledge-intensive industries, offering new business opportunities for workflow optimization and enhanced task efficiency.