OpenAI introduces GPT-6.1 Terra, pitching fast computer use for everyone

OpenAI has published an introduction page for GPT-6.1 Terra, a model it presents in a single line as bringing fast and capable computer use to everyone.

The framing is the substance. OpenAI did not lead with reasoning scores or context length but with computer use — the ability to operate software interfaces rather than only discuss them.

That positioning matters because computer use is the part of the agent stack hardest to make both quick and dependable: a model that is accurate but slow is useless for click-through work, and one that is fast but brittle breaks things.

The announcement is also light on the details buyers tend to ask for first. The page summary offers no benchmark table, no pricing and no rollout schedule, so Terra's practical profile remains to be established.

Read as a signal rather than a spec sheet, the release suggests OpenAI is treating desktop and browser control as a mainstream feature rather than a research demo; the phrase to everyone is doing real work in that sentence.

What to watch next: independent evaluations of how Terra handles long multi-step interface tasks, the safety controls around irreversible actions, and whether access arrives through the API, the consumer app, or both.