OpenAI introduced a powerful new AI model. It was prepared in "code red" mode

OpenAI has unveiled its most advanced artificial intelligence model, GPT-5.2, and it outperforms its predecessors in creating spreadsheets, preparing presentations, analyzing images, writing code and understanding long context, the company said. GPT-5.2 is available starting Dec. 11 for paid users of the ChatGPT chatbot and through a programmatic interface (API). With it, OpenAI hopes to attract more enterprise customers and increase revenue, Bloomberg reports .

The launch followed the recent unveiling of new models from rivals Anthropic and Google, prompting OpenAI to announce a "code red" mode - mobilizing resources to improve ChatGPT and temporarily freezing other projects, CNBC recalls . Google's Gemini 3 received high marks for reasoning and programming ability, quickly rose to the top of profile rankings and triggered a surge in the tech giant's stock, Bloomberg notes.As The Wall Street Journal wrote, the fact that it outperformed ChatGPT in benchmarks was of particular concern to OpenAI CEO Sam Altman, who said in an interview with CNBC on Thursday that the release of Gemini 3 had a smaller effect on the startup's performance than expected. Altman added that OpenAI will be out of code red mode by January, and "in a very strong position."

Google also announced a similar mode - shortly after launching ChatGPT three years ago, Bloomberg recalls.

Shares of Google's parent company Alphabet, which were declining by about 1.8% on the morning of December 11 amid a general sell-off in the AI sector after the disappointing Oracle report, reacted to the OpenAI presentation by strengthening the collapse and went down by 3%. They then trimmed some of the losses.

This article was AI-translated and verified by a human editor