Recent web traffic data from Similarweb reveals a growing challenge for OpenAI as Google’s Gemini continues to gain momentum in the AI race. While ChatGPT still leads in overall visitors, the gap is narrowing rapidly. In November, Google launched Gemini 3 Pro, the first major update in its Gemini 3 series, positioning it as a strong contender in the generative AI space. Since then, Gemini’s web traffic has surged by 28.4% month-over-month in December, while ChatGPT’s site traffic dropped by 5.6%. This shift, first noted by Menlo Ventures partner Deedy Das, underscores a significant change in user behavior. Although the data only tracks visits to chatgpt.com and gemini.google.com, it doesn’t account for usage through mobile apps, third-party integrations, or Google’s AI-powered search overviews—where the company holds a massive advantage. Year-over-year, the growth in traffic tells a more dramatic story. ChatGPT’s site traffic rose 49.5%, but Gemini’s jumped an astonishing 563.6%. In December, ChatGPT attracted 5.5 billion visitors, far outpacing Gemini’s 1.7 billion. However, the rate of growth suggests that Gemini is rapidly closing the gap, with other AI platforms like DeepSeek, Grok, Character.AI, Perplexity, and Claude all falling short of 400 million visitors. Gemini 3 was widely praised for its improved visual understanding, creative capabilities, and stronger coding performance—features that directly challenge ChatGPT’s core strengths. More importantly, Google’s ability to embed AI into its dominant search ecosystem gives it a unique edge. With billions of daily users already relying on Google Search, the company can seamlessly introduce AI features without requiring users to switch platforms. The competition extends beyond text. In the image generation space, Google’s Nano Banana Pro model launched just weeks before OpenAI unveiled ChatGPT Images, highlighting a fast-moving battle for market share. The data appears to have triggered a response at OpenAI. Internal communications revealed a “code red” alert, with CEO Sam Altman reportedly instructing teams to prioritize ChatGPT and delay other product initiatives. In a December appearance on the “Big Technology” podcast, Altman downplayed the immediate threat, saying the company wouldn’t remain in emergency mode for long and that “code red” periods typically last six to eight weeks. Still, he acknowledged that Gemini 3, much like DeepSeek before it, exposed weaknesses in OpenAI’s product strategy. “It did not have the impact we were worried it might,” Altman said, “but it did identify some weaknesses in our product offering strategy, and we're addressing those very quickly.”