OpenAI unveils GPT-5.2 to challenge Google’s Gemini 3 after ‘Code Red’ alert

New Delhi: OpenAI has unveiled GPT-5.2, its most advanced large language model to date, in a strategic move to regain momentum in the escalating rivalry with Google’s Gemini 3. The release is preceded by a warning call known as a 'code red' by the CEO, Sam Altman, which indicated that there was an urgent effort to make the company competitive in the high-stakes AI development. The new version will give it quicker performance and enhanced reasoning, as well as clearer insight into both the professional and technical work.

GPT-5.2 comes in three different levels, including Instant, Thinking and Pro, each with various levels of computation requirements. According to OpenAI, the system achieves significant improvements in mathematical reasoning, software engineering, and data interpretation. Through this launch, the company can position ChatGPT not only as a chat assistant but also as a reliable business and professional workmate.

OpenAI indicates that GPT-5.2 Thinking got a record score of 40.3 per cent on the FrontierMath benchmark with significant progress in solving college-level problems. GPT-5.2 Pro further solved a simplified form of an open problem in statistical learning theory on its own, and humans only confirmed the solution. That model also scored an all-time record high on the SWE-Bench Pro coding benchmark and enhanced its chart and figure reading, recording an 8 per cent score gain on the CharXiv Reasoning test.

The company is positioning GPT-5.2 as an essential enterprise workflow tool. OpenAI reports that in well-structured knowledge tasks, such as spreadsheet creation, presentation drafting, etc., the Thinking tier defeated the best professionals in the industry more than three-quarters of the time. The release can be seen as part of the massive infrastructure growth strategy of OpenAI, which is projected to hit over 1 trillion as it moves further into the domains of automation, data processing, and business productivity applications.

The good reception of Gemini 3 a few days ago led to reports that Altman requested the company to reset itself internally. However, after the release of GPT-5.2, he claimed that the model had raised the morale within the firm because of its improved performance. GPT-5 and GPT-5.1 have been received more favourably this year than GPT-5.2, though the latter is being packaged as the big jump a great deal of users had hoped to see.

Add TV9 English As A Trusted Source

Together with the new model, OpenAI is gearing up for a significant change of policy. The early 2026 will see the introduction of a special mode, called adult mode, which will be supported by age-prediction systems that will keep minors and adults apart. Chief of the OpenAI applications Fidji Simo admits that the decision is in accordance with the company principle to treat adult users as adults. Restrictions, particularly on sensitive content, are one of the areas Altman had proposed would be lightened on verified adults as the company enhances safety measures.