OpenAI launches GPT-5.2, its most advanced AI model yet. CEO Sam Altman says the company expects to exit “code red” mode by January.
OpenAI has unveiled GPT-5.2, its most advanced artificial intelligence model to date, marking a significant step in the company’s accelerated push to strengthen ChatGPT amid intensifying competition in the AI landscape.
Described as the best model yet for everyday professional use, GPT-5.2 delivers notable improvements across key tasks, including spreadsheet creation, presentation building, image perception, long-context understanding, and software development. The model is now available through ChatGPT and OpenAI’s API, expanding access for both individual users and enterprise developers.
The launch follows closely on the heels of GPT-5.1, released just weeks earlier, and comes during what OpenAI termed a company-wide “code red” initiative. The internal directive was introduced after rival firms, including Anthropic and Google, rolled out competing AI models, prompting OpenAI to reallocate resources and prioritize rapid innovation around ChatGPT.
Speaking to reporters, Fidji Simo, CEO of Applications at OpenAI, explained that the code red initiative was designed to sharpen focus and clarify priorities. While additional resources were directed toward ChatGPT, she emphasized that GPT-5.2’s release was the result of months of development rather than a reactionary move.
OpenAI CEO Sam Altman indicated that the competitive pressure—particularly following Google’s launch of Gemini 3—has been less disruptive than initially anticipated. He added that OpenAI expects to exit code red mode by January, signaling renewed confidence in the company’s trajectory.
“When a competitive threat emerges, the right approach is to focus intensely and resolve it quickly,” Altman noted, reinforcing OpenAI’s strategy of fast, targeted execution.
Three Versions for Distinct Use Cases
GPT-5.2 will be offered in three variants:
- Instant – optimized for speed and real-time information retrieval
- Thinking – designed for structured tasks such as planning and coding
- Pro – delivering the highest accuracy for complex and demanding queries
According to OpenAI, GPT-5.2 leads or matches top industry benchmarks, including SWE-Bench Pro for agentic coding performance and GPQA Diamond, which measures graduate-level scientific reasoning. On GDPval, OpenAI’s internal evaluation framework, the model outperformed or equaled top professionals on 70.9% of clearly defined tasks.
While Anthropic’s Opus 4.5 scores higher on SWE-Bench Verified, OpenAI maintains that its preferred benchmark, SWE-Bench Pro, offers greater resistance to data contamination and better reflects real-world industrial challenges.
A Decade of Transformation
Founded a decade ago as a research-focused organization, OpenAI has rapidly evolved into one of the world’s fastest-growing commercial AI companies following the launch of ChatGPT in 2022. Today, the chatbot is used by over 800 million people every week, underscoring OpenAI’s growing influence as AI becomes embedded in daily workflows across industries.
With GPT-5.2, OpenAI is reinforcing its ambition to lead the next phase of AI adoption—while making a strong case for its long-term vision amid a high-stakes race shaping the future of artificial intelligence.
Read More: Peru’s Top 10 Women Entrepreneurs to Watch in 2026
