OpenAI releases GPT-5.2 AI model in ChatGPT: Check what's new, availability

OpenAI's new GPT-5.2 model brings major upgrades in reasoning, coding, vision and long-context performance, with Instant, Thinking and Pro versions now rolling out to ChatGPT users

OpenAI has released its new GPT-5.2 AI model, bringing significant improvements across reasoning, coding, spreadsheet creation, presentation building, long-context understanding and advanced visual perception. The company says GPT-5.2 is its most capable system yet for handling complex, multi-stage professional workflows.

 

Alongside the model announcement, OpenAI confirmed that GPT-5.2 Instant, Thinking and Pro are beginning to roll out in ChatGPT for paid plans.

GPT-5.2: What’s new

OpenAI said that GPT-5.2 represents one of its “largest leaps yet”, introducing higher general intelligence, stronger long-context processing and more capable agentic tool-calling. The company added that the model performs more reliably on complex, multi-step tasks such as generating spreadsheets, building presentations, writing code, interpreting images and handling end-to-end real-world workflows.

A major highlight of the release is GPT-5.2 Thinking, which OpenAI calls its best model yet for professional use. OpenAI said that GPT-5.2 Thinking achieved state-of-the-art results in GDPval evaluation and became the company’s first model to perform at or above human-expert level.

 

Another key upgrade is long-context reasoning. OpenAI said GPT-5.2 can now maintain understanding and accuracy across hundreds of thousands of tokens, making it better suited for large documents such as technical reports, research papers, transcripts, contracts and multi-file codebases.

 

The company also described GPT-5.2 Thinking as its strongest vision model so far. OpenAI claims the model more accurately interprets dashboards, diagrams, UI layouts, product screenshots and other visually dense information. Its improved understanding of spatial relationships within images allows it to perform better on workflows in finance, engineering, design, operations and customer support. 

GPT-5.2 in ChatGPT

OpenAI said that users will immediately notice that GPT-5.2 is more structured, more reliable and more pleasant to interact with for day-to-day tasks. The new model powers three

versions inside ChatGPT: Instant, Thinking and Pro.

• GPT-5.2 Instant: Positioned as the fast, capable option, GPT-5.2 Instant brings improvements in information-seeking queries, how-to guidance, step-by-step explanations, technical writing and translation. It builds upon the more conversational tone introduced with GPT-5.1 Instant, while offering clearer and more concise answers that surface key information earlier.
• GPT-5.2 Thinking: Designed for deeper and more complex work, GPT-5.2 Thinking delivers more polished reasoning across coding, long-document summarisation, answering questions about uploaded files, solving maths or logic problems step-by-step and supporting planning and analytical decisions with better structure and richer detail.
• GPT-5.2 Pro: OpenAI described GPT-5.2 Pro as its smartest and most trustworthy model for difficult or high-stakes questions. Early tests show fewer critical errors and stronger, more consistent results in advanced fields such as software development and technical problem-solving.

GPT-5.2: Availability

OpenAI is rolling out GPT-5.2 Instant, Thinking and Pro in ChatGPT for paid plans — Plus, Pro, Go, Business and Enterprise. The company notes that the rollout will be gradual to maintain stability, so users who do not see the new models immediately are advised to check again later. The model is also available for developers in API.