GPT-6.1 Terra brings faster computer use to ChatGPT
By Steven Van ·
OpenAI’s GPT-6.1 Terra is a fast, lower-cost model for computer use, visual reasoning and writing, rolling out to paid ChatGPT plans and Codex.
OpenAI released GPT-6.1 Terra on October 26, a new model built for fast, capable computer use at a lower price. In ChatGPT, it is rolling out to Plus, Pro, Enterprise and Edu users in ChatGPT Work and Codex.
What it does
OpenAI says Terra understands what is on screen, takes action across apps and writes more naturally. ChatGPT pitches it as a model for the whole workday: it pairs sharper visual reasoning and precise computer use with clearer writing.
The examples OpenAI gives range from decoding technical drawings and drafting proposals to filling in forms and editing videos. The idea is that more of the routine work you do across apps can be handed to the model, at a cost that makes it practical to do often.
Availability
Terra is rolling out now to Plus, Pro, Enterprise and Edu users in ChatGPT Work and Codex. OpenAI also says it is coming to the API, where developers can use it to build apps and agents that read images, navigate software and take action. It is already listed on OpenRouter at $1 per million input tokens and $5 per million output tokens with a 1.05M-token context window, and GitHub made it generally available in Copilot the same day.
Why it matters
OpenAI is pitching Terra as fast computer use at a lower cost, suited to high-volume workflows. OpenRouter says it matches GPT-6 Astra on Agents’ Last Exam at one-tenth the price per token and finishes tasks more than three times faster on average. That makes handing routine screen work to a model cheaper to do often. OpenAI’s announcement does not say whether Free users will get Terra.
Sources: ChatGPT’s announcement and demo, OpenAI’s launch post, OpenAI Developers on API access and OpenRouter’s listing details.