OpenAI has officially released its GPT-5.2 model family, positioning it as the company’s most capable lineup yet for professional knowledge work. The launch arrives just weeks after internal concerns that OpenAI was losing ground to Google’s rapidly advancing Gemini models. GPT-5.2 is offered in three tiers: Instant for fast queries, Thinking for complex reasoning tasks, and Pro for maximum accuracy on difficult problems. Compared to GPT-5.1, version 5.2 shows measurable improvements across hallucination rates, vision understanding, coding ability, long-context reasoning, and tool usage. On GDPval, GPT-5.2 Thinking matched or outperformed industry professionals in 71% of real-world tasks, including spreadsheet analysis and presentation work. Despite internal requests to delay the release for additional polish, OpenAI pushed ahead as Gemini 3 began topping public leaderboards — signaling an increasingly competitive phase in frontier AI development.