OpenAI Group PBC Chief Executive Officer Sam Altman reportedly declared “code red” in a memo sent to the company’s employees on Monday.

The Wall Street Journal and The Information reported the internal announcement today. The publications’ sources also provided new details about OpenAI’s development roadmap. The company is reportedly working on a personal assistant, advertising features and a new flagship reasoning model.

“We are at a critical time for ChatGPT,” Altman reportedly wrote in the Monday memo. He informed staffers that OpenAI will work to improve the artificial intelligence service’s response speed, reliability and personalization options. The company also plans to broaden the range of questions that ChatGPT can answer.

Nick Turley, OpenAI’s head of ChatGPT, wrote on X that “our focus now is to keep making ChatGPT more capable, continue growing, and expand access around the world — while making it feel even more intuitive and personal.”

OpenAI reportedly plans to expedite the development effort by redirecting resources from other projects. Altman wrote that the company will delay those initiatives rather than scrap them. The Wall Street Journal reported that one of the projects relates to advertising, which suggests OpenAI may be planning to display ads in ChatGPT.

According to today’s reports, the company is also pausing work on a personal assistant dubbed Pulse. Additionally, the development freeze will affect projects focused on building AI agents for healthcare and shopping use cases.

The overhaul of OpenAI’s development roadmap comes about two weeks after Google LLC released Gemini 3, its new flagship large language model series. The model set a record on MathArena Apex, an AI benchmark that comprises particularly tough questions from recent math contests. Gemini 3 also outperformed competing LLMs across multiple programming benchmarks.

In conjunction with the model series’ launch, Google CEO Sundar Pichai disclosed that Gemini has about 650 million monthly active users. That’s a more than 40% increase from July. ChatGPT reportedly has over 800 million weekly active users.

OpenAI also faces increased competition from Anthropic PBC, which introduced a flagship LLM called Claude Opus 4.5 last week. The latter company says that its new model outperforms Gemini 3 across several coding and computer use benchmarks. Additionally, Anthropic stated that Claude Opus 4.5 is likely the safest LLM released by any developers. 

According to the Journal, OpenAI is gearing up to release a new reasoning model of its own that can likewise best Gemini 3 in some areas. The paper’s sources didn’t specify the tasks at which the upcoming LLM is better. OpenAI is expected to launch the model next week.

Photo: Focal Foto/Flickr