At the beginning of this month, CEO of OpenAI, Sam Altman, announced a ‘code red’ following Google’s unveiling of Gemini 3. Reports indicate that the AI company has postponed certain plans to focus more intensely on ChatGPT, particularly with the upcoming release of GPT-5.2. This situation suggested that Altman’s organization was undergoing a period of heightened tension. Nevertheless, the OpenAI leader asserted that this was not the first instance of a code red scenario for the firm.
Sam Altman is ready to be vigilant during code red situations
In a recent episode of the Big Technology Podcast, Sam Altman stated that OpenAI has encountered “code red” alerts on several occasions. These alerts, which arise in response to major advancements by competitors, have become a standard strategy for the organization. Altman emphasized the importance of acting swiftly when faced with competitive threats, asserting, “It’s beneficial to be paranoid and respond quickly when a potential competitor arises.”
Deepseek also initiated a code red at OpenAI
The OpenAI CEO observed that while Google Gemini 3’s influence was not overly significant, it served to highlight certain areas where improvements were necessary for his company. He suggested that this situation resembled the launch of Deepseek. Altman explained, “[Gemini 3] did not have the repercussions we feared it might. However, it did — much like Deepseek — expose some vulnerabilities in our product strategy, and we are addressing those promptly.”
Altman noted that prior to the release of Gemini 3, OpenAI had entered code red after Deepseek’s launch.
Sam Altman anticipates that code red alerts will remain a consistent aspect of OpenAI’s operations, suggesting that they enable the company to tackle weaknesses relative to its competitors. He remarked, “I suspect we will continue to have these alerts maybe once or twice each year for a long time, as it is crucial for ensuring our success in the industry.”
OpenAI expected to exit code red soon
In response, OpenAI has swiftly implemented various upgrades. The organization has recently launched GPT-5.2, an enhanced AI model aimed at improving ChatGPT’s capabilities in professional, coding, and scientific applications, alongside a significant update for ChatGPT Images.
Reflecting on the typical length and frequency of these emergency situations, Altman stated, “Historically, these have generally lasted about six to eight weeks for us.” Currently, the OpenAI leader believes that the company will not remain in code red for “much longer.”
