OpenAI is currently in emergency mode. Last week, CEO Sam Altman is said to have announced a “Code Red” within the company, which is an internal term used for the highest level of urgency. According to reports, this alert was issued after Google’s Gemini 3 model became increasingly popular. Launched last month, Gemini 3 is Google’s fastest and most capable AI model to date, and it even surpasses OpenAI’s current ChatGPT in multimodality, logic, reasoning, efficiency, speed, and more.

Multiple reports, including an internal memo reviewed by The Wall Street Journal, suggest that OpenAI is now focused on prioritizing the release of its next model, GPT-5.2, with teams being reallocated, features put on hold, and timelines accelerated as the company works to keep ChatGPT ahead of the AI ​​race.

Here’s 5 points you need to know about the ongoing emergency inside OpenAI headquarters.

Point 1: Sam Altman declared “Code Red” inside OpenAI

Sam Altman issued a “Code Red” order last week and directed the entire organization to mobilize to improve ChatGPT’s core capabilities, according to reports citing people familiar with internal discussions. Inside OpenAI, this label is used sparingly and only for moments that require immediate redirection of resources.

And this directive doesn’t just indicate urgency. Reportedly, OpenAI employees have been asked to pause non-essential work and refocus on strengthening the models at the heart of the platform. The order is being treated as a company-wide rallying point, shifting focus away from long-term experiments toward short-term, higher-priority competitive goals.

Point 2: What caused the Emergency?

The trigger reportedly came in November when Google released Gemini 3, which is believed to be the most powerful AI model Google has ever created. This new model from DeepMind Labs excels in deep multimodal reasoning across text, media, and code, and uses autonomous agentic workflows to deliver significantly improved accuracy in complex technical domains. The model also demonstrated strong reasoning, coding, and multimodel capabilities.

Following its launch, Gemini 3 topped several AI benchmark charts, drawing comments from leading tech giants and AI leaders including OpenAI CEO Sam Altman and XAI CEO Elon Musk.

Meanwhile, for OpenAI, this performance gap created by Gemini 3 was significant enough to raise real concerns that Google has overtaken ChatGPT in predictive ability. The Gemini app, which had about 450 million monthly active users in July 2025, reportedly grew to more than 650 million soon after. This surge in popularity increased competitive pressure, prompting urgent internal analysis and ultimately receiving a Code Red order from OpenAI bosses.

Point 3: OpenAI fast-tracks the launch of GPT-5.2

As part of the Code Red situation, OpenAI has reportedly accelerated the release of GPT-5.2, the next major update powering ChatGPT. According to reports, this model was originally targeted for release in late December. However, now the update is expected to happen around December 9th based on the final stability check.

As reported by The Information, internal evaluation shows that OpenAI’s next reasoning model is already outperforming Gemini 3 in early internal tests. These results have reportedly increased confidence within the company, leading to the decision to release GPT-5.2 earlier than planned in an effort to regain its lead in the AI ​​race.

Point 4: Many ChatGPT features are being stopped

Meanwhile, to support accelerated work on GPT-5.2, OpenAI has reportedly stopped or slowed down development on several ChatGPT features that were planned for an upcoming update. These include experimental tools, interface redesigns, and creative add-ons that do not directly affect the core behavior of the model.

Instead the teams have been reassigned to tasks focused on improving ChatGPT’s stability, response time, logic accuracy, and overall reliability. OpenAI’s engineering efforts are currently aimed at ensuring that users see a clear performance boost once GPT-5.2 is introduced.

Point 5: OpenAI is reorganizing product schedules and internal workflows

Under the Code Red directive, OpenAI has begun to adjust project schedules and reorganize internal workflows to match the new pace of development. According to reports, OpenAI Workdesk teams are now working in shorter, more intensive development cycles, with roadmaps being updated more frequently than usual.

This includes moving personnel between departments, reassigning engineers to core model initiatives, and strengthening coordination between research, product, and infrastructure teams. OpenAI is reportedly considering GPT-5.2 as the first milestone in this accelerated phase, with additional model updates and refinements already being planned under the revised architecture.