Model available through ChatGPT Work, Codex and API
OpenAI has introduced GPT-6.1 Terra, a new AI model designed to combine speed and efficiency with detailed visual reasoning, precise computer use and a clearer, more natural writing style.
On Agents’ Last Exam, a benchmark evaluating AI agents’ ability to use software, GPT-6.1 Terra scored 60% at the maximum reasoning effort level, at a cost of $1.01. By comparison, GPT-5.6 Terra scored 50.7% at $3.58, while Anthropic’s Claude Opus 5.5 scored 63.2% at $8.82.
In OpenAI’s internal tests, GPT-6.1 Terra’s error rate at low reasoning effort fell to 16.7%, down from 53.6% for GPT-5.6 Terra. At maximum reasoning effort, the error rate dropped to 4.3%, which the company says is close to the performance of GPT-6 Astra and GPT-6.1 Sol.
GPT-6.1 Terra is available to Plus, Pro, Enterprise and Edu subscribers through ChatGPT Work and Codex. Developers can also access it through the OpenAI API.
API pricing:
- Input tokens: $1 per 1 million tokens.
- Cached input tokens: $0.05 per 1 million tokens.
Note: The source text does not specify the price for output tokens.
