OpenAI Launches GPT-5.2: Discover the Enhancements Over 5.1!

In a recent press release from its media center, OpenAI announced the launch of its new GPT-5.2 model series, designed “to enable everyone to create even more economic value,” as stated by the company.

GPT-5.2: OpenAI’s New Model Series

Among the model series released by OpenAI, GPT-5.2 is described as “the most advanced to date for intellectual work of professionals,” according to the press release. This model series targets a broad audience, including professionals, and has been a focus of OpenAI for over a year.

True to its tradition, OpenAI has introduced three versions:

• GPT-5.2 Thinking: This is the most advanced model for professionals, optimized for complex, high-value tasks. This version shows notable progress in programming, deep analysis, long document synthesis, mathematical and logical reasoning, as well as financial modeling, planning, and structured decision-making.
• GPT-5.2 Instant: A fast and versatile model for everyday work and learning, especially effective for information searches, practical guides, technical writing, translation, and support in studies and career orientation, with improved accuracy.
• GPT-5.2 Pro: The most reliable and high-performing model for challenging questions, designed to maximize answer quality at the cost of longer computation times, with fewer major errors and strong performance in complex areas like programming.

GPT-5.2: Enhancements Over GPT-5.1

What improvements does GPT-5.2 bring over its predecessor, GPT-5.1? OpenAI highlights seven areas of significant enhancement. Based on internal benchmarks, these improvements confirm that GPT-5.2 (particularly the Thinking version) outperforms GPT-5.1 in the following areas:

1. High-value tasks: GPT-5.2 Thinking significantly enhances its ability to perform complex intellectual tasks (such as presentations, spreadsheets, reports) with quality comparable to human experts, being three times faster and considerably less costly.
2. Programming: GPT-5.2 Thinking now excels at debugging production code and implementing features, allowing developers to correct and refactor complex codebases with much less manual intervention. It notably surpasses previous versions in front-end development.
3. Veracity: Reducing model hallucinations has been a challenge for OpenAI, and GPT-5.1 was no exception. With GPT-5.2 Thinking, the company promises a 38% reduction in hallucinations compared to its predecessor, translating to responses with 38% fewer errors in research, writing, analysis, and decision support.
4. Extended context: GPT-5.2 Thinking now handles very long documents with much more accuracy and consistency, enabling professionals to analyze contracts, reports, and complete articles without losing critical information.
5. Vision: GPT-5.2 Thinking roughly halves error rates in reasoning from graphs and understanding software interfaces, enabling more accurate interpretations of dashboards, screenshots, and technical diagrams to support workflows in finance, operations, engineering, design, and customer support.
6. Tool calls: GPT-5.2 Thinking manages complex multi-step workflows more efficiently, facilitating the orchestration of data retrieval, analysis, and deliverable generation without execution breaks.
7. Science and mathematics: GPT-5.2 Pro and GPT-5.2 Thinking assist researchers in tackling more complex scientific and mathematical problems, accelerating progress in both fundamental and applied research.

We designed GPT‑5.2 to help everyone create even more economic value. It performs better in creating spreadsheets, designing presentations, writing code, analyzing images, understanding lengthy contexts, using tools, and managing complex projects in multiple stages.

Available in ChatGPT and the API

GPT‑5.2 is being rolled out today in ChatGPT for users of paid plans (Plus, Pro, Business, Enterprise). The deployment is being done gradually to ensure service stability. GPT‑5.1 will remain available in ChatGPT for three months under the section for older models.

In the API, GPT‑5.2 is immediately accessible under the identifiers gpt-5.2, gpt-5.2-chat-latest, and gpt-5.2-pro. Pricing is set at $1.75 per million input tokens and $14 per million output tokens. GPT‑5.2 Pro costs $21 for input and $168 for output.