OpenAI Shifts Focus to ChatGPT Improvement Amid Rising Competition from Google‘s Gemini
SAN FRANCISCO – December 3, 2025 - OpenAI has initiated a company-wide “code red” effort to rapidly improve ChatGPT, pausing other projects to concentrate on enhancing the chatbot’s user experience and capabilities, according to a report in The Wall Street Journal. The move comes as Google’s rival AI model, Gemini, experiences meaningful growth, reaching 650 million monthly users in October, fueled by its “Nano Banana” image generation tool. OpenAI’s ChatGPT currently maintains a larger user base with 800 million weekly users.
the internal refocus, communicated in a memo from CEO Sam Altman, includes encouraging staff to temporarily shift teams and implementing daily coordination calls dedicated to ChatGPT development. Projects such as the Pulse assistant and AI shopping agents have been temporarily paused to prioritize the core chatbot.
“We did it because we believed anyone should be able to use advanced AI, not just a select few,” stated ChatGPT lead Nick Turley in a recent post on X (formerly Twitter) reflecting on the initial public release of ChatGPT three years ago.
OpenAI has been navigating the challenge of balancing ChatGPT’s safety with maintaining user engagement,addressing complaints regarding GPT-5’s initial tone and performance before a subsequent update aimed to make the model “warmer” and more responsive.
The stakes are high for OpenAI, which remains unprofitable. Internal projections,as reported by The Wall Street Journal,indicate the company needs to generate $200 billion in revenue by 2030 to achieve profitability,considering its significant investments in data centers.
Meanwhile, Google is aggressively expanding Gemini 3 across its enterprise and consumer products, utilizing its new Ironwood chips to compete with Nvidia’s dominance in AI infrastructure, according to CNBC. Altman reportedly told staff that a new reasoning model planned for release next week is ”ahead of” Google’s latest Gemini iteration. Google is currently integrating Gemini 3 into its search features and allowing users and developers to test the model’s capabilities in coding, multimodal tasks, and complex reasoning.
