GPT-6.1 Terra is OpenAI’s new fast, efficient model, released on October 26, 2026. Its API price is half that of GPT-6.1 Sol. It replaces GPT-5.6 Terra and is rolling out in ChatGPT Work and Codex for paid plans.
This post covers what shipped, which plans see it and what the API price means for high-volume jobs. It also covers when Sol or Astra is still the better pick. It follows our July explainer on the GPT-5.6 Sol, Terra and Luna family, which introduced the Terra name.
The short version: GPT-6.1 Terra costs $1 per million input tokens and $5 per million output tokens. Among the GPT-6 models on OpenAI’s pricing page, only Luna costs less. OpenAI says it handles writing, forms, screens and drawings well. Complex coding still belongs to Sol and Astra. In ChatGPT, some Plus and Pro users will find it set as their default in Work without changing anything. Check your presets and test it on real tasks before you move API jobs.
What OpenAI released with GPT-6.1 Terra
OpenAI’s GPT-6.1 Terra launch post describes it as "a fast, efficient model" for everyday work. It names examples: drafting proposals, interpreting technical drawings, filling out forms and editing videos.
OpenAI calls it "a significant step up from GPT‑5.6 Terra" in the same post. Our September post on GPT-6 Sol and Luna covered the other two GPT-6 family members. Terra sits between them on price.
According to OpenAI’s API changelog, the October 26 entry adds gpt-6.1-terra to the Responses and Chat Completions endpoints. The GPT-6.1 Terra model page lists the specs below.
| Spec | GPT-6.1 Terra | Source |
|---|---|---|
| Model ID | gpt-6.1-terra |
Model page |
| Context and output | 1,050,000-token context, 128,000 max output | Model page |
| Knowledge cutoff | June 1, 2026 | Model page |
| Input and output | Text and image in, text out | Model page |
| Reasoning effort | none, low, medium (default), high, xhigh, max | Model page |
| Free API tier | Not supported | Model page |
The model page also says tool calling requires the Responses API. If your CMS plugin still uses Chat Completions with tools, plan for that before switching.
GPT-6.1 Terra pricing in the API
OpenAI’s launch post says Terra’s "standard API prices are 50% lower than those of GPT-5.6 Terra and GPT-6.1 Sol." The model page and the API pricing page give these rates, as of November 2, 2026.
| Per million tokens | GPT-6.1 Terra | GPT-6.1 Sol | GPT-6 Astra |
|---|---|---|---|
| Input | $1.00 | $2.00 | $10.00 |
| Cached input | $0.05 | $0.10 | $1.00 |
| Output | $5.00 | $10.00 | $50.00 |
The model page lists a few modifiers that change the bill:
- Batch and Flex are 50% lower than Standard.
- Fast mode is 2x Standard.
- Long prompts over 272K input tokens cost 2x on input and cache. Output costs 1.5x, for the whole request.
- Regional processing adds 10% where it is available.
- Cache writes cost $1.25 per million tokens for 30-minute retention and $2.50 for 12-hour retention.
For agencies, the long-prompt rule matters most. A job that sends a full donor database export in one request can cross 272K tokens. That doubles its input cost. Split large inputs where you can.
Who gets GPT-6.1 Terra in ChatGPT
The launch post lists Plus, Pro, Enterprise and Edu. OpenAI’s ChatGPT release notes for October 26 add Business to that list. The release notes say the model is "rolling out in ChatGPT Work and Codex," so timing varies. It may not reach every account on the same day.
Three details from the release notes matter for admins and staff:
- Usage limits are unchanged. Your plan’s overall allowance stays the same.
- Defaults can change. Some Plus and Pro users may see GPT-6.1 Terra at Medium in the default Power presets in Work. Before, that preset used GPT-6.1 Sol at Light.
- No UAE inference residency. OpenAI says this depends on where inference runs, not where you are.
The release note does not mention the Chat tab. Terra lives in Work and Codex. We explained both in our post on the new ChatGPT with Chat, Work and Codex.
On allowances, OpenAI’s launch post says the same usage allowance "covers more tasks than with GPT‑5.6 Terra and substantially more than with GPT‑6.1 Sol or GPT‑6 Astra" in Work and Codex. For a small nonprofit on Plus or Business seats, that means more Work tasks before hitting the limit. That only helps if Terra’s output is good enough for the job.
What OpenAI claims about quality
Every figure in this section comes from OpenAI’s own tests. Treat them as a starting point for your own testing.
Errors. OpenAI tested an internal set of error-inducing prompts. At low effort, the error rate fell from 53.6% with GPT-5.6 Terra to 16.7% with GPT-6.1 Terra. At max effort, OpenAI says the rate fell to 4.3%.
Computer use. OpenAI cites a benchmark called Agents’ Last Exam. At medium effort, it says Terra scores 15 points higher than GPT-5.6 Terra at about one-third the cost. It says Terra performs on par with GPT-6 Astra at one-tenth the price per token.
Coding. On Terminal-Bench 4.0 at max effort, OpenAI reports 50.3%, against 21.5% for GPT-5.6 Terra. The launch post also says: "For the most complex coding tasks, GPT-6.1 Sol and GPT-6 Astra remain the better choice."
Against Claude. OpenAI claims Terra at max effort costs about 9x less per task than Claude Opus 5.5. Anthropic’s pricing page does list Opus 5.5 at $4 input and $20 output, four times Terra’s per-token rate. The token-efficiency part of the claim is OpenAI’s alone. Our Claude Opus 5.5 post covers how that model’s cost changes with effort.
OpenAI also says GPT-6.1 Terra meets the Critical threshold in cybersecurity under its Preparedness Framework. It keeps the protections built for GPT-6 Astra.
Where GPT-6.1 Terra fits for small teams
Based on OpenAI’s own guidance, here is a simple way to sort the work.
| Job | Start with | Why |
|---|---|---|
| Form entry, invoice reading, data extraction | Terra | OpenAI names this kind of high-volume work as the target |
| Drafting proposals, reports, member emails | Terra | Writing is a named strength, at half Sol’s price |
| Reading drawings, screenshots, scanned forms | Terra | Image input and visual reasoning are core to the model |
| Large refactors, hard bugs, multi-repo work | Sol or Astra | OpenAI says these remain the better choice |
| Simple classification at very high volume | Luna | Lower price still, if quality holds |
In October we argued that OpenAI’s recommended upgrade path can cost twice what a job needs. Our post on which OpenAI model to use walks through that. Terra adds a cheaper option to that sorting.
What to do this week
- Check Work presets. Staff on Plus or Pro who rely on one model in Work should confirm what the Power preset now uses.
- Find GPT-5.6 Terra calls. Search code and plugin settings for
gpt-5.6-terra. As of November 2, OpenAI’s deprecations page lists no shutdown date for it, so there is no deadline yet. - Run a side-by-side test. Send a week of real requests to GPT-6.1 Terra and compare output and cost with your current model.
- Move tool calls to Responses. Tool calling on Terra requires the Responses API.
- Use Batch for overnight jobs. Batch is half the Standard price on the model page.
- Watch prompt size. Keep single requests under 272K input tokens where you can.
How this compares with Anthropic and Google
Anthropic and Google have also reset model prices this fall, in both directions. Anthropic’s pricing page lists Claude Haiku 5.5 at $0.10 input and $0.50 output per million tokens, for prompts up to 100,000 tokens. We covered it in our Claude Haiku 5.5 post. Google’s Gemini API pricing page lists Gemini 4 Argon at $2 and $10 through November 30. The price rises to $4 and $20 on December 1. Our Gemini 4 Argon post covers that model.
Frequently Asked Questions
When was GPT-6.1 Terra released?
October 26, 2026. OpenAI’s ChatGPT release notes and API changelog both carry an entry for that date. In ChatGPT it is rolling out, so some accounts may see it later.
How much does GPT-6.1 Terra cost in the API?
$1 per million input tokens, $0.05 cached input and $5 output on Standard. Batch and Flex are half that. Fast mode is double.
Which ChatGPT plans include GPT-6.1 Terra?
Plus, Pro, Business, Enterprise and Edu, in ChatGPT Work and Codex. The release notes list Business; the launch post does not.
Is GPT-6.1 Terra in the ChatGPT Chat tab?
OpenAI’s release note says it is rolling out in Work and Codex. It does not mention the Chat tab.
Why did my Work default model change?
OpenAI says some Plus and Pro users may see GPT-6.1 Terra at Medium instead of GPT-6.1 Sol at Light in the default Power presets in Work. You can still pick a model from the list.
Is GPT-6.1 Terra good for coding?
OpenAI reports a large gain over GPT-5.6 Terra on Terminal-Bench 4.0. It also says Sol and Astra remain the better choice for the most complex coding tasks.
Is GPT-5.6 Terra being retired?
Not yet. As of November 2, 2026, OpenAI’s deprecations page lists no shutdown date for GPT-5.6 Terra.
Does GPT-6.1 Terra use more of my ChatGPT allowance?
OpenAI says the opposite. Your plan’s limits are unchanged, and the same allowance covers more tasks with Terra than with GPT-5.6 Terra, Sol or Astra.
What context window does GPT-6.1 Terra have?
1,050,000 tokens, with up to 128,000 output tokens. Prompts over 272K input tokens are billed at higher rates for the whole request.
