GPT-6.1 Terra is OpenAI's fast model for images, computer use and writing, at half GPT-6.1 Sol's price

OpenAI has released GPT-6.1 Terra, a fast, efficient model for visual reasoning, precise computer use, and clear, natural writing. It's in the API today, through the Responses and Chat Completions endpoints.

Slide text

01 02 03 What it's for

OpenAI API changelog · Oct 26, 2026 · OpenAI API docs · GPT-6.1 Terra model page

OpenAI releases GPT-6.1 Terra

Visual reasoning Computer use Natural writing

What it's for.

Terra takes text and images in and writes text out. Use it to match product photos to catalog entries, compare technical drawings, act on application screens, or draft and revise text.

Terra fills a new slot in the GPT-6 family. Astra is for the hardest work, Sol balances speed and cost, Luna is the fastest and cheapest, and Terra sits between Sol and Luna.

It holds about a million tokens of context and writes up to 128,000 tokens in one reply. Its knowledge runs to June 1st, 2026, a month later than GPT-6.1 Sol's.

What it costs.

Terra costs one dollar per million input tokens and five dollars per million output tokens. That's half the price of GPT-6.1 Sol. Luna stays ten times cheaper for high-volume jobs.

Unlike Astra and GPT-6.1 Sol, Terra can switch reasoning off entirely. Start at medium, the default, and raise it when a better answer is worth the extra time and tokens.

Very long prompts change the bill. Past 272,000 input tokens, the whole request is billed at twice the input rate and one and a half times the output rate.

Where it runs.

Terra is an API model, and the free tier can't call it. Rate limits start at 5,000 requests a minute on the Build tier and rise to 15,000 on Grow.

In the Responses API, Terra works with OpenAI's built-in tools. That covers computer use for acting on screens, web and file search, code interpreter, a hosted shell, and MCP servers.

Chat Completions serves Terra too, but only without tools. If your app calls tools through Chat Completions today, those requests have to move to the Responses API to run on Terra.

For quicker answers, Fast mode runs Terra at twice the standard price. It supports EU data residency for Terra, which GPT-6 Astra and GPT-6 Sol don't offer in Fast mode.

If you run GPT-5.6 Terra on screenshots or product photos, I'd try GPT-6.1 Terra first, because that's the work it's built for. Compare both on your own images at the same reasoning effort. OpenAI's Using GPT-6 guide covers the switch.