OpenAI
GPT-6.1 Terra
visual reasoning, computer use and writing at $1 input and $5 output

Key facts
- 26 Oct 2026OpenAI API changelog
- Released
- $1 / $5per million input / output tokens
- API price
- $0.05per million tokens, standard rate
- Cached input
- 1,050,000tokens across the request
- Context
- 128,000tokens
- Maximum output
- Text + imagestext output
- Input
GPT-6.1 Terra is an AI model from OpenAI for understanding images, working with computer interfaces and writing. Developers can use it through the API, and eligible subscribers can choose it in ChatGPT Work and Codex. Its standard API prices are half those of GPT-6.1 Sol.
GPT-6.1 Terra is an AI model that can interpret pictures, write text and use software through connected tools. OpenAI released it on 26 October 2026, with a particular focus on visual reasoning and computer interfaces.
OpenAI’s examples include matching product photographs to catalogue entries, reading technical drawings and checking how an image has been cropped. Acting on a computer also requires the relevant tools, access and permissions. A model’s ability to understand a screen is one part of that workflow.
Terra accepts text and images
The model documentation lists a 1,050,000-token context window, which is the space available for the request and response. Maximum input is 922,000 tokens and maximum output is 128,000. Terra produces text, with a knowledge cut-off of 1 June 2026.
Developers call gpt-6.1-terra. Both Chat Completions and Responses support the model; tool calling requires the Responses API. Reasoning settings range from none through low, medium, high and xhigh to max, with medium as the default. These settings control how much reasoning the model performs before answering.
Standard API prices start at $1 and $5
Terra’s standard API rates are half the input and output prices of GPT-6.1 Sol.
| Per million tokens | Standard rate | Request with more than 272,000 input tokens |
|---|---|---|
| Input | $1.00 | $2.00 |
| Cached input | $0.05 | $0.10 |
| Cache write, 30-minute retention | $1.25 | $2.50 |
| Cache write, 12-hour retention | $2.50 | $5.00 |
| Output | $5.00 | $7.50 |
The higher long-context rates apply to the whole request. Batch and Flex rates are 50% below Standard, while Fast mode costs twice Standard. Regional processing adds 10% where offered. Connected tools can carry their own charges, so compare the complete task cost as well as the token rates.
Paid subscribers can choose Terra in Work and Codex
OpenAI’s release update lists Plus, Pro, Business, Enterprise and Edu access in ChatGPT Work and Codex. Enterprise availability also depends on workspace controls. API usage is billed separately from a ChatGPT subscription.
Terra’s ChatGPT availability is through Work. Its published residency options exclude UAE inference residency, a condition about where processing runs. Browser and desktop actions have their own setup and access requirements.
Choose a model around the work it needs to do
OpenAI positions Terra for visual tasks and writing, GPT-6.1 Sol for more complex coding and professional work, GPT-6 Astra for its most demanding tasks, and GPT-6 Luna for focused, high-volume jobs. Those are the provider’s product roles. Testing the same representative task across models gives a more useful comparison for a particular application.
Questions people ask
- What does GPT-6.1 Terra do?
- GPT-6.1 Terra accepts text and images and produces text. With the appropriate tools and permissions, it can work with browser and desktop interfaces. OpenAI positions it for visual reasoning, computer use and writing.
- How much does GPT-6.1 Terra cost?
- Standard API rates are $1 per million input tokens, $0.05 for cached input and $5 for output. Requests with more than 272,000 input tokens use higher rates for the entire request. ChatGPT subscriptions and API billing are separate.
- Where can I use GPT-6.1 Terra?
- The API model ID is gpt-6.1-terra. OpenAI also offers it in ChatGPT Work and Codex on Plus, Pro, Business, Enterprise and Edu, subject to workspace availability and controls.
More in Large Language Models
All LLMs →- OpenAIGPT-6.1 Solthe upgraded mid-price GPT-6 model, released 29 September 2026
- OpenAIGPT-6 Astrathe new generation, rolling out from 3 September 2026
- OpenAIGPT-6 Sol and Lunathe mid-price and budget GPT-6 models, released 22 September 2026
- Moonshot AIKimi K3.1 Proa reasoning model for coding, 3D design and agent tasks
- Google DeepMindGemini 4 Argonlonger reasoning, coding and cyber defence
- OpenAIGPT-5.6the flagship from 9 July to 3 September 2026