OpenAI releases GPT-5.5, taking the company one step closer to becoming an AI “super app”

OpenAI on Thursday released its latest AI model, GPT-5.5, which the company calls its “smartest and most intuitive model” to date. The algorithm has enhancements in a number of areas, and OpenAI co-founder and president Greg Brockman claims it brings the company one step closer to creating OpenAI “super apps.”

In a call with journalists, Brockman said the new model is a major step forward “toward more agentic and intuitive computing.”

“This model is a real step toward the kind of computing we expect in the future. But it’s one step, and we expect to see many more in the future,” Brockman said. “Compared to something like 5.4, we can do faster, sharper thinking with fewer tokens. So this means more frontier AI is available to businesses and consumers, and that’s part of our goal.”

Brockman also said the model is an additional step toward creating a “super app” — a multipurpose Swiss Army knife of programs — that Brockman and co-founder Sam Altman had previously talked about launching. The co-founders envision combining ChatGPT, Codex, and AI browsers into one integrated service that can help enterprise customers. Notably, the “super app” concept has also been talked about by Altman’s rival (and former OpenAI colleague) Elon Musk, who has said he wants to turn X into its own so-called super app.

OpenAI just released its last model last month, with the previous release being in December and before that in November. The company continues to churn out new models at a rapid pace, and company staff say this trend is expected to continue for some time. “We’re seeing pretty significant improvements in the short term and very significant improvements in the medium term,” said Jakub Pachocki, principal scientist at OpenAI. “Actually, I think the last two years have been surprisingly slow.”

According to OpenAI, GPT-5.5 is designed to be useful in a wide range of categories, including basic enterprise areas such as agenttic coding and knowledge work, as well as more experimental AI applications such as mathematics and scientific research. The company also released data Thursday showing the model’s strong performance across a variety of benchmarks. According to OpenAI, the 5.5 consistently scores higher than previous models and models from competitors Google and Anthropic, such as the Gemini 3.1 Pro and Claude Opus 4.5.

The rivalry between OpenAI and Anthropic has been contentious, and GPT-5.5 provided a new opportunity for both companies to compare themselves to each other. One of the reporters during the press conference asked about GPT 5.5. It will have similar functionality to Anthropic’s recently announced cybersecurity tool Mythos. (Mythos has experienced controversy in recent days due to reports of unauthorized access to its program.) Mia Glaese, a member of OpenAI’s technical staff, said GPT-5.5 will have a significant impact on the company’s approach to deploying models for digital defense. “We have a strong long-standing strategy for our approach to cyber and have refined a durable approach to safely deploying our models,” Glaese said.

Mark Chen, chief research officer at OpenAI, said that GPT-5.5 is better at handling computer work than previous versions, and also said that the model “represents meaningful advances in scientific and technical research workflows,” noting that the company feels the model can truly “help advance professional scientists.” Chen said it could also help in the field of drug discovery, which has seen increased industry interest in recent years.

OpenAI says GPT 5.5 will be widely available starting Thursday. The company says this model is being rolled out to Plus, Pro, Business, and Enterprise users of ChatGPT, and 5.5 Pro is aimed at Pro, Business, and Enterprise users.