OpenAI CEO Sam Altman has admitted that increasing competition in the AI ​​field is putting pressure on the company, forcing it to move faster and rethink priorities more often than before. Speaking on a recent episode of the Big Technology Podcast, Altman said that OpenAI has entered “code red” mode internally several times, triggered by rival AI models threatening to close the gap with ChatGPT. According to him, this increased state of urgency is no longer an exception, but something the company hopes to revisit regularly as the race for AI leadership intensifies.

Altman explained that OpenAI deliberately responds aggressively whenever a potential competitive threat emerges. He said being hypervigilant and responding quickly has become part of the company’s operating mindset.

“It’s good to be nervous and take immediate action when a potential competitive threat emerges,” he said, adding that such emergency steps may happen once or twice every year in the near future, as OpenAI focuses on staying ahead in a fast-moving market.

One such “code red” moment earlier this year was triggered by the sudden rise of China-based DeepSeek. The company surprised the global tech industry in January by claiming that its AI model could match leading systems, including OpenAI’s own o1 model, while operating at much lower costs. Altman said this development prompted OpenAI to reevaluate its approach and respond immediately.

Recently, soon after Google launched its latest AI chatbot, Gemini 3, OpenAI entered another Code Red phase. Although Gemini 3 was released in November, it has attracted significant attention in recent weeks, with Google touting it as its most advanced AI model to date. About two weeks after its release, Altman reportedly informed OpenAI employees via an internal Slack message that ChatGPT would become the company’s top priority, even if it meant delaying other product plans.

Despite initial concerns, Altman clarified that Gemini 3 did not cause as much disruption as OpenAI had anticipated. “It didn’t have the impact that we were worried it might have,” he said on the podcast. However, he acknowledged that, like DeepSeek, Google’s new model exposed some shortcomings in OpenAI’s broader product strategy. Altman said these vulnerabilities have already been identified and are being addressed rapidly. He did not disclose the features that the company is preparing to fix.

But since it entered its latest Code Red phase, OpenAI has released several updates. The company recently introduced a more capable AI model designed to boost ChatGPT’s performance in professional tasks, including coding and scientific work. Earlier this week, OpenAI also introduced a new image-generation model, suggesting a renewed focus on improving both core features and creative tools within its ecosystem.

Altman said the current state of emergency is not expected to last much longer. Based on past experience, he said such intensive internal phases typically last six to eight weeks before the company returns to normal operations.

But, the idea of ​​a “Code Red” response is not unique to OpenAI. In fact, Google itself declared an internal Code Red in 2022 after the public launch of ChatGPT. At the time, Google was seen as lagging behind in consumer-facing AI products, despite playing a major role in funding and developing the research that laid the foundation for today’s AI boom.