OpenAI PBC CEO Sam Altman issued an internal "red alert" in a memo to staff on Monday, according to reports.

Both The Wall Street Journal and The Information published details of the announcement today, citing sources familiar with OpenAI’s updated development roadmap. The company is reportedly accelerating work on a personal assistant, ad-supported features, and a new flagship reasoning model.

"We are at a pivotal moment for ChatGPT," Altman wrote in his Monday memo. He emphasized efforts to enhance the AI service's speed, reliability, and personalization, as well as expanding the range of questions ChatGPT can answer.

Nick Turley, head of ChatGPT at OpenAI, added on X: "Our focus right now is making ChatGPT more capable, driving continued growth, and expanding global access—while also making it feel more intuitive and personalized."

To expedite development, OpenAI plans to reallocate resources from other initiatives rather than cancel them outright. According to The Wall Street Journal, one such delayed project involves advertising, hinting at potential ad integration into ChatGPT.

The company has also paused work on Pulse, its personal assistant initiative, and imposed a development freeze on projects focused on AI agents for healthcare and shopping use cases.

This strategic overhaul comes roughly two weeks after Google launched its new flagship large language model series, Gemini 3. The model set a record on MathArena Apex—a benchmark featuring especially challenging problems from recent math competitions—and outperformed rival LLMs across multiple coding benchmarks.

Concurrently with the model launch, Google CEO Sundar Pichai revealed that Gemini now boasts approximately 650 million monthly active users, a more than 40% increase since July. Meanwhile, ChatGPT reportedly exceeds 800 million weekly active users.

OpenAI also faces intensifying competition from Anthropic PBC, which last week unveiled its new flagship model, Claude Opus 4.5. Anthropic claims the model surpasses Gemini 3 in several coding and computer-use benchmarks and describes it as potentially the safest LLM ever released by any developer.

According to The Wall Street Journal, OpenAI is preparing to launch its own new reasoning model, which reportedly outperforms Gemini 3 in certain areas. While sources did not specify the exact tasks where the upcoming LLM excels, the model is expected to be released next week.