GPT-6.1 Terra
GPT-6.1 Terra is OpenAI’s October 26, 2026 model for visual reasoning, computer use and writing. It accepts text and images and returns text. API developers use gpt-6.1-terra; tools require the Responses API, while ChatGPT Work and Codex have their own plan and rollout conditions.
Read the official documentation ↗Why it matters
OpenAI model for visual analysis, computer use and writing with lower standard token prices than GPT-6.1 Sol.
What Terra can help you do
Use Terra to interpret a technical drawing, match a product photograph to a catalog entry, work through an application screen, or draft and revise a document. A computer-use workflow also needs the application, tools, permissions and acceptance checks; the model does not independently provide your desktop or authorize actions.
API contract and full-request costs
The API lists 1,050,000 context tokens, 128,000 maximum output tokens and a June 1, 2026 knowledge cutoff. Reasoning effort supports none, low, medium (default), high, xhigh and max. Standard rates per million tokens are $1 input, $0.05 cached input and $5 output. Cache writes cost $1.25 for 30-minute retention or $2.50 for 12 hours. Above 272K input tokens, the whole request uses 2× input/cache rates and 1.5× output rates. Fast is 2× Standard; Batch and Flex are separate half-price modes.
- GPT-6.1 Terra API record
Exact endpoint, input, price and unexecuted request example.
ChatGPT access is a separate contract
OpenAI’s launch lists Plus, Pro, Enterprise and Edu access in Work and Codex, subject to rollout, region and workspace settings. This does not establish ordinary Chat-tab access. Terra does not support UAE inference residency; that restriction concerns the inference location, not the user’s physical location.
How to compare Terra with Sol or Astra
OpenAI reports stronger visual and computer-use results than GPT-5.6 Terra under named benchmark and effort settings. Those are vendor evaluations, not evidence of the same speed, quality or cost on your task. Compare the same images, tools, effort and verified outcome; lower per-token prices do not guarantee lower total task cost. OpenAI still points to Sol or Astra for the hardest coding and reasoning work.
- GPT-6.1 Sol
Compare the documented model and workload boundary.
- Official Terra demonstrations
Watch the publisher’s visual and computer-use examples.
Useful first jobs
- Visual analysis of documents and application screens.
- Supervised computer tasks with explicit tools and permissions.
- Drafting and revising business documents.
Frequently asked questions
Can GPT-6.1 Terra generate images or videos natively?
Its documented native output is text, with text and image input. A Responses API workflow may call an image-generation tool; that does not turn Terra into a native image or video output model.
Does Terra cost half as much as Sol?
Terra’s standard input and output token rates are half GPT-6.1 Sol’s standard rates. Actual task cost still depends on input length, cache writes, output and reasoning tokens, service tier, tool calls and retries.
Sources and evidence
Current publisher documentation controls the documented contract. Publisher examples and evaluations are attributed; GetLLMs has not executed the model or installed the software for this review.
- GPT-6.1 Terra API model contract ↗
OpenAI · Primary source
- Introducing GPT-6.1 Terra ↗
OpenAI · Primary source
- ChatGPT release notes ↗
OpenAI · Primary source
- Terra system-card addendum ↗
OpenAI · Primary source
- Terra discussion ↗
Hacker News · Community experience