OpenAI CEO Sam Altman recently declared a “code red” effort to immediately improve ChatGPT. “We are at a critical time for ChatGPT,” Altman said in news reports based on an internal memo. “I expect the vibes out there will be rough for a bit.”
In corporate terms, a “code red” signals an existential threat—issues that could incinerate a business if left unaddressed. OpenAI’s urgency appears justified. In the two weeks following Gemini 3’s launch, ChatGPT’s daily active users dropped six percent, according to Deedy Das, a former Google engineer and current venture capitalist. Average daily visits plummeted from 203 million to 191 million—a decline suggesting a major competitive shift.
The response has been dramatic. OpenAI has paused work on advertising integration, AI shopping agents, and PULSE, a personalized digital assistant meant to compete with Siri and Google Assistant. Instead, the company is laser-focused on upgrading ChatGPT’s platform stability, reducing latency, and addressing user complaints that the model has become “cold,” less creative, and prone to errors.
From Meteoric Rise to Market Pressure
ChatGPT’s trajectory has been remarkable. While Instagram took 2.5 years to reach 100 million users and TikTok needed nine months, ChatGPT hit an estimated 100 million monthly active users within just two months of launch. Its combination of a large pre-trained transformer model (GPT-3.5) and simple conversational interface revolutionized how people interact with AI.
But the AI landscape shifts quickly. While chatbots have existed since ELIZA in 1966, and popular assistants like Siri (2011) and Alexa (2014) have become household names, the current generation of AI models represents a quantum leap in capability—and competition.
Google Strikes Back
Google’s November release of Gemini 3, seamlessly integrated across its ecosystem, has changed the game. The model has earned accolades on benchmark leaderboards like LMSYS Chatbot Arena for text generation and multimodal reasoning. According to Forbes, Google’s AI user base has surged to 650 million monthly users from 450 million in August following the release of its image generator, Nano Banana Pro.
Meanwhile, Anthropic released Claude Opus 4.5 last month, claiming it as the world’s best coding model. The pressure on OpenAI is mounting from multiple directions.
The contrast in financial positions is stark. Google generates billions in profit annually, while OpenAI—despite its $500 billion valuation—projects $20 billion in revenue for 2025 but burned through $2.5 billion in just the first half of last year. The company faces AI infrastructure commitments potentially exceeding $1.4 trillion over the next decade, with no clear path to profitability before 2030.
Enter Garlic
OpenAI’s answer is Garlic, a new large language model—possibly GPT-5.2 or GPT-5.5—potentially arriving in early 2026. Its mission: surpass rivals in advanced reasoning and coding abilities, the exact areas where Gemini 3 currently dominates.
Mark Chen, OpenAI’s Chief Research Officer, has publicly touted Garlic’s “strong performance” in internal testing. With the “code red” declaration, all company resources have shifted toward Garlic’s development and launch.
“We know we have some work to do,” Altman admits, “but we are catching up fast.”
The Pendulum Swings
There’s irony in OpenAI’s predicament. Just three years ago, Google declared its own “code red” in response to ChatGPT’s sudden market dominance. But Google’s vast resources and revenue allowed it to weather that storm and strike back decisively. Can OpenAI, despite its massive valuation, do the same without profitability?
The Garlic model represents more than a performance upgrade—it’s a financial defense designed to stop the hemorrhaging of users and revenue. With promises of massive cost-efficiency and superior reasoning, it must not only retain users but transform billions in red ink into black.
Will this technological remedy or dose of Garlic be potent enough? Time will tell.
