OpenAI fires back at Google with GPT-5.2 after ‘code red’ memo

On Thursday, OpenAI introduced its latest model, GPT-5.2, positioning it as the most advanced iteration yet, tailored for both developers and everyday professional use. This model is now available to ChatGPT's paid users and developers through an API, offering three distinct versions: Instant, designed for quick responses to common queries such as writing and translation; Thinking, which specializes in complex tasks including coding and document analysis; and Pro, the premium option that ensures maximum accuracy for challenging problems. "We crafted GPT-5.2 to enhance economic value for users," stated Fidji Simo, OpenAI’s Chief Product Officer, during a media briefing. He highlighted the model's improved capabilities in creating spreadsheets, crafting presentations, and understanding complex visual and contextual information. This launch comes as OpenAI faces stiff competition from Google’s Gemini 3, which has been leading benchmark evaluations in most categories, with the exception of coding, where Anthropic’s Claude Opus-4.5 remains dominant. Earlier this month, reports surfaced about an internal 'code red' memo from CEO Sam Altman, expressing concerns over declining ChatGPT traffic and the potential loss of market share to Google. The memo prompted a shift in focus towards enhancing the ChatGPT experience rather than pursuing plans for advertising. The introduction of GPT-5.2 marks OpenAI’s strategic move to regain its competitive edge, even as some team members suggested delaying the release for further improvements. Despite indications of an increased focus on consumer personalization, the launch seems to prioritize enterprise applications, targeting developers and the broader tooling ecosystem to establish itself as the primary platform for AI-driven applications. Recent data from OpenAI indicates a significant rise in enterprise usage of its AI tools over the past year. This surge coincides with Google’s rollout of managed MCP servers, facilitating easier integration of its services like Maps and BigQuery with AI systems. OpenAI claims that GPT-5.2 sets new performance benchmarks in various areas, including coding, mathematics, science, and reasoning, which could facilitate more reliable workflows and complex systems capable of handling extensive data. During the briefing, OpenAI's Product Lead Max Schwarzer emphasized that GPT-5.2 offers substantial enhancements in code generation and debugging, with reported performance improvements noted by coding startups like Windsurf and CharlieCode. Additionally, the Thinking variant of GPT-5.2 has shown a 38% reduction in errors compared to its predecessor, enhancing its reliability for everyday tasks and decision-making. While GPT-5.2 builds upon the advancements of its predecessors, it also represents a consolidation of previous updates. The model is designed to be a more dependable foundation for production use, especially as OpenAI commits to significant investments in AI infrastructure to support its growth amidst intensifying competition from Google. However, the focus on reasoning capabilities poses financial risks for OpenAI, as these advanced systems demand more computing resources, potentially leading to high operational costs. Reports suggest that OpenAI's expenditures on computing resources may be greater than previously disclosed, complicating its financial landscape. Interestingly, the launch did not include any updates on image generation capabilities, despite Altman's acknowledgment of its importance in the code red memo. Following Google's successful release of the Nano Banana image model, which gained viral attention, OpenAI is expected to announce a new model in January that will offer improved image generation and performance. In addition, the company is implementing new safety measures concerning mental health usage and age verification for teens, although these changes were not a focal point during the launch event.