OpenAI rolls out GPT-5.2 soon after declaring 'code red' warning
What's the story
Just a month after the launch of GPT-5.1, OpenAI has unveiled the next-gen model, GPT-5.2, for its popular chatbot service. The new model is touted as "the most capable model series yet for professional knowledge work." It is designed to help users complete tasks more quickly and efficiently across a range of activities like spreadsheet creation, presentation building, coding, image perception, and multi-step project completion.
Model improvements
Enhanced capabilities and performance
The new model comes with enhanced general intelligence, long-context understanding, agentic tool-calling, and vision. These improvements make it ideal for real-world professional use. Compared to its predecessor GPT-5.1, the thinking capabilities of GPT-5.2 have improved significantly with 30% fewer errors in responses. It can now handle reports, contracts, papers and multi-file projects while maintaining accuracy across hundreds of thousands of tokens.
Benchmark success
Performance surpasses human experts
OpenAI claims that GPT-5.2 outperforms industry professionals in knowledge work tasks across 44 occupations, scoring 70.9% on the GDPval test. This is a huge leap from its predecessor, which scored only 38.8%. It is also the first model from OpenAI to perform at or above a human expert level, marking a major milestone in AI development.
User experience
New features and user experience
The latest model will offer a more structured and reliable experience for ChatGPT users, with a warmer, more conversational tone. It also comes with three modes: Instant, Thinking, and Pro. While the Instant mode is designed for everyday tasks like info-seeking questions and technical writing, the Thinking mode caters to complex tasks such as summarizing long documents or coding. The Pro mode is best suited for difficult questions where a high-quality answer is worth waiting for.
Rollout details
Rollout and future plans
The new model is now available in ChatGPT for paid users, with the API accessible to all developers. This launch comes just a week after OpenAI CEO Sam Altman issued a "code red," calling on employees to prioritize improving ChatGPT in light of competition from Google's Gemini and Anthropic's Claude. Despite the competitive pressure, Altman downplayed the impact of Gemini 3's launch on his company.