OpenAI has released GPT-5.2, the latest update to ChatGPT, signaling a continued shift toward more reliable, work-ready artificial intelligence. The model was introduced on December 11, 2025, following an announcement earlier in the week, and is now being integrated directly into ChatGPT for users across multiple plans.
The rollout began with paid subscribers, including Plus, Pro, Go, Business, and Enterprise users, with access for free users expanding gradually. Rather than introducing flashy new features, GPT-5.2 focuses on under-the-hood improvements designed to make the system more dependable in everyday and professional use.
GPT-5.2 is available in three variants: Instant for fast, everyday interactions, Thinking for deeper reasoning and multi-step analysis, and Pro for advanced and sustained workloads. OpenAI says the model delivers stronger reasoning, improved long-context handling, and fewer factual errors compared with GPT-5.1, particularly in tasks that require careful analysis or research.
Performance improvements are also evident in how the model behaves. GPT-5.2 responds more smoothly, with faster output and reduced lag, making interactions feel more fluid and responsive. More notably, OpenAI says the model demonstrates measurable gains on internal benchmarks tied to knowledge-work tasks, with performance approaching — and in some cases exceeding — human-level results in specific professional scenarios, including research, analysis, and structured reasoning.
Without making sweeping claims, the benchmark results suggest a narrowing gap between human expertise and AI-assisted work — a shift with growing implications for how professionals research, decide, and create.
