OpenAI has unveiled its latest language model, GPT-5.2, at a time of heightened competition in the AI field. The company says this new version aims to “create more economic value for people.” Applications CEO Fidji Simo told reporters at a briefing on Thursday that the model has been under development for “many, many months.”
GPT-5.2 Unveiled
The company states that GPT-5.2 is “the best model to date for everyday professional use.” This move marks a clear competition against Gemini 3, which has gained a reputation as a general-purpose primary model.
The GPT-5.2 model series includes Instant, Thinking, and Pro models. The company says this series is particularly successful in “spreadsheet creation, presentation preparation, coding, visual perception, understanding long contexts, using tools, and managing complex, multi-step projects.”
OpenAI’s vice president of research, Aidan Clark, stated in a briefing that they had given a senior immunology researcher access to GPT-5.2 Pro. Clark noted that the researcher asked the model to generate the most important unanswered questions about the immune system, and that the researcher reported that the model produced “sharper questions and stronger explanations of why those questions are important” compared to other leading models.
In its blog post, OpenAI also stated that GPT-5.2 is better suited for AI agent workflows. This is seen as part of the battle among AI companies to deliver the most efficient and useful AI agent. Researcher Max Schwarzer said in the briefing that their vision is to make ChatGPT the best “personalized assistant.”
The company also reported that the Thinking model “experiences fewer hallucinations” than the previous model. They specifically emphasized that this difference makes it more useful for professionals seeking reliable AI agent tools. Among the test users of the model before its release are companies like Notion, Box, Shopify, Harvey, Zoom, and Databricks, which gained access a few weeks ago.
The release of GPT-5.2 comes less than a month after the release of GPT-5.1, where OpenAI largely focused on making the models “warmer” and “more enjoyable to talk to.” The new model builds on these tonal changes: “Users should find that GPT-5.2 feels better in everyday use, more structured, more reliable, and still enjoyable to talk to,” OpenAI says in a blog post. With this release, the company is also responding to Google’s recent advancements in artificial intelligence.
Last week, OpenAI CEO Sam Altman declared a “code red” within the company, indicating that it would be postponing certain initiatives, such as advertising, to support improvements to ChatGPT. Simo didn’t share additional details about ad timelines at Thursday’s briefing, but said, “we announced this code in red to signal to the company to aggregate resources in this particular area.”
The GPT-5.2 announcement came on the same day news broke that OpenAI has signed a three-year licensing agreement with Disney to enable “user-requested social videos” featuring over 200 characters from Disney, Marvel, Pixar, and Star Wars. As part of the deal, Disney will stream some of these videos on Disney Plus and become a “major customer” of OpenAI, along with a $1 billion equity investment in the AI company.
