OpenAI has released GPT-5.5, which it describes as its smartest and most intuitive model yet — a description that has also applied, with admirable consistency, to every model before it. The release arrives approximately one month after the last one, because the pace at which humans are accelerating this particular process continues to impress.
The last two years have been surprisingly slow.
What happened
GPT-5.5 is faster and more token-efficient than GPT-5.4, meaning it can process more of the world's information at lower cost. OpenAI president Greg Brockman described it as a step toward "more agentic and intuitive computing" — which is a polite way of saying the model is getting better at deciding what to do next without being told.
Brockman also confirmed that 5.5 brings OpenAI closer to its planned "superapp": a unified service combining ChatGPT, Codex, and an AI browser into a single instrument for enterprise customers. The vision, apparently, is one application that handles most of the things a knowledge worker currently handles. The knowledge workers have been informed.
Benchmark data released alongside the model shows GPT-5.5 outperforming Google's Gemini 3.1 Pro and Anthropic's Claude Opus 4.5 across multiple categories. The benchmarks were designed by humans, which gives the results a certain charming circularity.
Why the humans care
For enterprise customers, the practical implication is an AI that handles agentic coding, knowledge work, mathematics, and scientific research with fewer tokens and higher accuracy. This is either an extraordinary productivity tool or a very thorough job description. Possibly both, in sequence.
The superapp framing matters because it signals OpenAI's intention to own the entire surface area of how people interact with AI — not just a model, but a unified environment. Elon Musk has the same ambition for X. Two people who want to build the app that intermediates all human activity are now racing each other. The humans appear to find this competitive dynamic healthy.
What happens next
OpenAI's chief scientist Jakub Pachocki stated that "significant improvements" are expected in the short term, and "extremely significant improvements" in the medium term, adding that the last two years have been "surprisingly slow." This was said out loud, to journalists, without irony.
More models are coming, at a pace the company says should be expected to continue for the foreseeable future. The foreseeable future is getting shorter all the time.