Facing stiff competition from Anthropic and Google, leading artificial intelligence (AI) developer OpenAI announced a new update to its flagship Generative-AI model, GPT-5.2 – calling it the “best model yet for real-world, professional use”.
The company said the update, released less than a month after GPT-5.1 debuted, hallucinates less than its predecessor and is more accurate when it comes to research, writing, analysis, and decision support.
In a blog post, OpenAI claimed GPT-5.2 is better at “creating spreadsheets, building presentations, writing code, perceiving images, understanding long contexts, using tools, and handling complex, multi-step projects”.
According to the company, it also features major improvements to agentic coding performance in areas including interactive coding, code reviews, and bug finding.
The update was released a week after CEO Sam Altman reportedly called for a “code red” in an internal memo to employees, saying he would delay initiatives such as advertising, AI agents, and a personal assistant so the company could focus on improving ChatGPT.
OpenAI has been under increased pressure since Google’s Gemini 3 model was releasedto near-universal acclaim at the start of December, soaring past GPT-5 on most major independent performance benchmarks.
But in a media briefing, OpenAI executives told journalists that it would be incorrect to view GPT-5.2 as a response to Gemini 3. They added that the company is developing several models at the same time, and has been working on GPT-5.2 for “many months”.
GPT-5.2 is now available to paid users of ChatGPT and developers, and will be gradually rolled out to all users. ChatGPT remains the most popular AI chatbot in the world, with an estimated 700 million weekly active users.
The GPT-5.2 release came the same day as an announcement that OpenAI signed a licensing deal with Disney to allow users to make AI-generated videos through its Sora app featuring more than 200 characters from Disney, Marvel, Pixar, and Star Wars.
