GPT-6.1 Terra: OpenAI's cheaper computer-use model, price and what's new
OpenAI's GPT-6.1 Terra is a faster, cheaper model for computer use, priced at $1 per million input tokens and $5 per million output tokens. Here is what changed, what it costs and when to pick it over Sol.
By ScoviAI Team
Key points
- Released October 26, 2026 as gpt-6.1-terra; available in ChatGPT Work and Codex for Plus, Pro, Enterprise and Edu users, and via the API
- API price: $1 per million input tokens and $5 per million output tokens, half of GPT-6.1 Sol's $2 and $10
- OpenAI-reported (internal test, via Decrypt): factual error rate at low effort 16.7% vs 53.6% for GPT-5.6 Terra; Terminal-Bench 4.0 at max effort 50.3% vs 21.5% for GPT-5.6 Terra
- 1,050,000-token context window; prompts over 272K input tokens cost 2x input and 1.5x output
- OpenAI still points to Sol or Astra for complex coding and in-depth research
OpenAI has released GPT-6.1 Terra, a faster and cheaper model built for "computer use": an AI that looks at your screen and clicks and types the way a person would. On the API it costs $1 per million input tokens and $5 per million output tokens, half the price of GPT-6.1 Sol, and OpenAI says standard prices are 50% lower than the previous Terra, GPT-5.6 Terra.
What is GPT-6.1 Terra?
Terra is the middle model in OpenAI's lineup. OpenAI's model page describes it as "a fast, efficient model for visual reasoning, precise computer use, and clear, natural writing," and suggests uses such as matching product photos to catalog entries, interpreting technical drawings, acting on application screens, and drafting or revising text.
The model ID is gpt-6.1-terra. It takes text and images as input and returns text. It has a 1,050,000-token context window, up to 128,000 output tokens and a June 1, 2026 knowledge cutoff. Reasoning effort can be set to none, low, medium (the default), high, xhigh or max. Tool calling requires the Responses API.
How much does GPT-6.1 Terra cost?
According to the GPT-6.1 Terra and GPT-6.1 Sol pages, standard text prices per million tokens are:
| Model | Input | Output |
|---|---|---|
| GPT-6.1 Terra | $1 | $5 |
| GPT-6.1 Sol | $2 | $10 |
As an illustration (our arithmetic, not OpenAI's): a job that uses 10 million input tokens and 1 million output tokens would cost about $15 on Terra and about $30 on Sol at standard rates. Two details matter for real bills. Prompts longer than 272,000 input tokens are charged at 2x the input rate and 1.5x the output rate for the whole request, and OpenAI's pricing notes say tool-specific products such as search and computer use can carry per-call fees, so check the pricing page for your setup. Batch and Flex are 50% cheaper, while Fast mode costs twice the standard price.
What can it do better than before?
OpenAI says the new model is "a significant step up over GPT-5.6 Terra, with 50% lower standard API prices," a line quoted in Decrypt's coverage of the launch. Decrypt also reports the following figures, which it attributes to OpenAI's own testing (we could not open OpenAI's launch post directly, so these are as reported by Decrypt):
- On an internal test of prompts designed to trick models into factual mistakes, the error rate at low reasoning effort was 53.6% for GPT-5.6 Terra and 16.7% for the new Terra (4.3% at maximum effort).
- On Terminal-Bench 4.0, which measures how many command-line tasks a coding agent completes, Terra at maximum effort solved 50.3% of tasks versus 21.5% for GPT-5.6 Terra.
- OpenAI says Terra finishes tasks more than three times faster on average than GPT-6 Astra and GPT-6.1 Sol.
These are vendor-reported numbers on partly internal tests, so treat them as a direction, not a guarantee. Independent benchmarks have not been published yet in the sources we reviewed.
Who can use it?
Per Decrypt's report of the announcement, Terra is rolling out to Plus, Pro, Enterprise and Edu users in ChatGPT Work and in Codex, and developers can call it through the API. For in-depth research or complex coding, OpenAI still points users to GPT-6.1 Sol or GPT-6 Astra, so Terra is not meant to replace them. For background on the family, see our earlier coverage of GPT-6.1 Sol and how GPT-6 reached every ChatGPT user.
What does GPT-6.1 Terra mean for you?
- If you use ChatGPT on a paid plan: you should see Terra as an option for routine screen-based chores, such as filling in forms or moving data between apps. Check the model picker, since rollouts reach accounts gradually.
- If you build apps: Terra is the cheaper option to test for agent workflows. Compare it with Sol on your own tasks at the same reasoning effort before switching, and check the pricing page for possible per-tool-call fees and the long-prompt surcharge.
- If you are cautious about AI acting on your computer: an agent that can click through your apps can also make mistakes in them. Start with low-stakes tasks and keep a person in the loop for anything involving money or sensitive data.
For more on what OpenAI announced to developers this month, read our DevDay 2026 recap.
Sources
- GPT-6.1 Terra model page (OpenAI)
- GPT-6.1 Sol model page (OpenAI)
- OpenAI Launches GPT-6.1 Terra (Decrypt)
ScoviAI Team is the ScoviAI newsroom. We write every article ourselves from the original announcements and link our sources. Spotted a mistake? Tell us and we'll fix it.
ADVERTISEMENT
