OpenAI launches GPT-6.1 Terra for lower-cost visual reasoning and computer use

OpenAI has released GPT-6.1 Terra, a GPT-6 family model positioned for visual reasoning, computer use, coding, and high-volume professional work. Developers can call it as gpt-6.1-terra in the API. OpenAI also says Plus, Pro, Enterprise, and Edu users can use it in ChatGPT Work and Codex. GitHub separately announced that it is available in GitHub Copilot.

For cost-sensitive teams, the launch post lists standard API prices of $1 per million input tokens, $0.05 per million cached input tokens, and $5 per million output tokens. OpenAI links the model to jobs such as document interpretation, form completion, technical-drawing work, and video editing, and reports gains against GPT-5.6 Terra on its evaluations. Those benchmark results are vendor claims and are not a substitute for an independent evaluation on a production workflow.

This is a new model release, not a feature update to one existing tool. It gives model-routing teams another option to compare for capability, latency, and cost in visual-input and controlled computer-use tasks. Start in a low-risk, reversible environment, then measure end-to-end completion, human interventions, total token spend, and the effect of your safety boundaries before expanding traffic.

Computer use still needs least-privilege credentials, explicit confirmations, and an audit trail. A model's alignment description or vendor benchmark does not guarantee safe behavior in every browser, desktop, file format, or internal system.

View the GPT-6.1 Terra catalog entry