OpenAI released GPT-5.2 to fully hit Google (GOOG.US, GOOGL.US) Gemini 3 to accelerate the seizure of the enterprise-level AI market

Just a few weeks after Google (GOOG.US, GOOGL.US) launched the acclaimed Gemini 3, and OpenAI was positively challenged, the AI unicorn quickly made a strong response, and a new artificial intelligence model, GPT-5.2, was officially launched to comprehensively enhance ChatGPT's capabilities in programming, scientific research, and complex work tasks.

The Zhitong Finance App learned that OpenAI said that GPT-5.2 is faster, information retrieval is more accurate, and there are significant improvements in writing and translation. The new model is divided into three versions, and also enhances the ability to simulate human reasoning processes, and is designed to handle more complex and time-consuming tasks such as mathematics, science, and programming.

As Google and Anthropic launch new models over the past few weeks, OpenAI's leading edge in the AI competition has been hit like never before. Google Gemini 3, in particular, quickly topped several authoritative industry rankings due to its outstanding reasoning and coding performance, including Lmarena and Humanity's Last Exam.

Faced with competitive pressure, OpenAI CEO Ultraman announced a “red alert” (code red) in an internal memorandum at the beginning of this month and requested all departments of the company to concentrate resources and make every effort to accelerate the iterative upgrade of ChatGPT.

With the launch of GPT-5.2, OpenAI hopes to attract more enterprise customers and increase revenue to support its infrastructure investment plans of more than $1 trillion over the next few decades. The company said that the new model has significant enhancements in processing Excel tables, creating presentations, and writing code, with the goal of “creating higher economic value for users.” GPT-5.2 was the first to be rolled out to paid users starting Thursday.

In August of this year, OpenAI launched the highly anticipated GPT-5, but market reviews were mixed, and some users believed that the increase fell short of expectations. GPT-5.1, which was then launched in November, was soon taken away by the more powerful Gemini 3. Ultraman said in an interview yesterday that “Gemini 3's impact on our metrics is lower than expected,” and he expects OpenAI to “step out of the red alert mode with a strong attitude” by January next year.

On the same day, OpenAI also revealed that it has begun to launch software that can predict the age of users to better manage how ChatGPT presents content to minors. Fidji Simo, head of OpenAI's application department, said the company plans to launch the much-publicized “adult model” next year.