In response to competitive pressure from Google’s Gemini 3, which has surpassed AI standards and garnered considerable industry acclaim since its November launch, OpenAI plans to accelerate its timeframe and release its GPT-5.2 model on Tuesday, December 9.
On December 1, CEO Sam Altman issued a “code red” for the entire company, which is the highest alert level on OpenAI’s three-point scale. He urged staff to concentrate on enhancing ChatGPT’s speed, dependability, and customization while postponing other projects like advertising, shopping tools, and a personal assistant called Pulse. The Verge claims that GPT-5.2 is prepared for release and should “close the gap that Google created” with the introduction of Gemini 3.
The AI Race Is Upended By Google’s Gemini 3
The urgency comes after Google’s Gemini 3 launch on November 18, which dominated several benchmarks and attained a record-breaking score of 1501 Elo on the LMArena Leaderboard. The model achieved 91.9% on GPQA Diamond compared to GPT-5.1’s 88.1% and 37.2% on Humanity’s Last Exam compared to GPT-5.1’s 26.5%.
After testing Gemini 3, Salesforce CEO Marc Benioff openly gave up on ChatGPT, saying on X: “I’ve used ChatGPT every day for three years.” Gemini 3 was used for two hours just now. I will not return. Reasoning, speed, graphics, video—everything is faster and sharper—the leap is insane. Elon Musk, the CEO of xAI, and Sam Altman both commended Gemini 3’s potential.
According to SimilarWeb data, ChatGPT lost about 6% of its traffic in late November, falling from 106 million to 100 million daily active users while Gemini’s traffic increased by 40%. While OpenAI maintains over 800 million weekly active users, Google revealed that Gemini achieved 650 million monthly active users in October, up from 450 million in July.
Financial Stress Increases
The pressure from competitors coincides with OpenAI’s growing financial difficulties. In the first half of 2025, the corporation made $4.3 billion in sales but suffered $7.8 billion in operating losses, mostly from ChatGPT operations and research expenses. OpenAI anticipates sustained losses until 2028 despite anticipating yearly sales of over $20 billion.
Altman warned staff that GPT-5.2 tested “ahead” of Gemini 3 in internal evaluations and noted that the corporation should be ready for “rough vibes” in his internal memo. As OpenAI competes with Google and Anthropic, which released its Claude Opus 4.5 model on November 24, the update will concentrate core enhancements over additional capabilities.