GPT-6.1 Terra API

NewLLM

GPT-6.1 Terra is the cost-optimised GPT-6.1 route: 1M context, image input and tool use, billed at exactly half of GPT-6.1 Sol.

Read the docs Get your API key

Input · USD / 1M tokens$0.1125 USDOfficial price~~$1 USD~~
Output · USD / 1M tokens$0.675 USDOfficial price~~$5 USD~~
Savings91%
Context1M
View tool pricing and billing rules

What are we making today?

Bring an idea, a question, or a first draft.

What this model is for

GPT-6.1 Terra is the cost-optimised GPT-6.1 route: 1M context, image input and tool use, billed at exactly half of GPT-6.1 Sol.

GPT-6.1 Terra is OpenAI's cost-focused model in the GPT-6.1 family, released on October 26, 2026 as gpt-6.1-terra. OpenAI places it in the third slot of the lineup, behind GPT-6 Astra and GPT-6.1 Sol and above GPT-6 Luna, and describes it as built for visual reasoning, precise computer use and clear, natural writing. The catalog lists a 1M context window, image input and tool use, and bills it below GPT-6.1 Sol on input, output and cache reads.

Choose Terra for volume: reading invoices into accounting tables, drafting proposals from templates, or clicking through a desktop application. Unlike Sol it accepts the none effort setting, so classification and short rewrites can skip reasoning tokens altogether, and its knowledge cutoff of June 1, 2026 sits a month later. Pair it with Sol when a task needs the deepest coding pass, and keep Terra as the default for the long tail.

Pick GPT-6.1 Terra for high-volume GPT-6.1 work that does not need Sol's reasoning depth; every rate is half of Sol's.

Best-fit coding workloads

Invoice field extraction

Read scanned or photographed invoices and return supplier, date, line items and tax as JSON. Image input and the 1M window let a stack of pages travel in one request, and the none effort setting keeps short extractions from spending reasoning tokens. Validate totals against the document before writing to the accounting system.

Technical drawing to spec

Turn architectural or mechanical drawings into a written specification: dimensions, tolerances, materials and open questions. Send each sheet as an image and ask for a field-by-field list that cites the location on the sheet, then have a person confirm ambiguous dimensions before they reach production.

Form filling on screen

Automate repetitive desktop forms: send a screenshot, take the next click or field value, and repeat. Tool use on the Responses API carries the actions, and one task per conversation keeps a wrong entry cheap to undo. Ask the model to quote the visible label it acted on.

Hands-on notes

Checked on 2026-10-27 through the OmniaKey API with one standard coding prompt (median of 3 runs) and three capability probes.

Output tokens for our standard coding prompt, reasoning included: 196
Cost of that output at current prices: $0.0001
JSON mode: Passed
Tool calling: Passed
Image input: Passed

Technical specification

Catalog facts and integration details verified for this route.

Model ID: gpt-6.1-terra
Context: 1M
Inputs and tools: Vision · Tool use
API endpoint: /v1/chat/completions
Cache hit / 1M: $0.01125 USD
Coding clients: Codex · Cursor · Cline · Aider

Limits to plan around

The 128,000-token output ceiling is shared with Sol and Astra, so a request that returns a book-length document still needs splitting.
OpenAI names Sol and Astra as the better choice for the most complex coding, so a full agent rewrite is not Terra's job; escalate that pass to Sol.
Tool calling and computer use need the Responses API. Chat Completions remains for requests without tools.
The knowledge cutoff of June 1, 2026 predates recent library releases, so new APIs have to be pasted into the prompt rather than assumed.

Compare nearby models

Compare workload fit and current listed token rates before choosing a route.

Model	Best for	Context	Input	Output
GPT-6.1 TerraThis page gpt-6.1-terra	The volume default in GPT-6.1: forms, drawings and desktop steps where a cheaper pass is enough.	1M	$0.1125 USD Official price $1 USD	$0.675 USD Official price $5 USD
GPT-6.1 Sol gpt-6.1-sol	Pick Sol for the hardest coding and long investigations; Terra lists lower rates in every token category.	1M	$0.225 USD Official price $2 USD	$1.35 USD Official price $10 USD
GPT-5.6 Terra gpt-5.6-terra	The predecessor lists higher input, output and cache rates, so new work belongs on the 6.1 model.	1M	$0.3 USD Official price $2 USD	$1.8 USD Official price $12 USD

SDK quickstarts

Use one base URL and model ID from cURL, Python, or TypeScript.

Streaming chat completion

curl https://api.omniakey.com/v1/chat/completions \
  -H "Authorization: Bearer $OMNIAKEY_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gpt-6.1-terra",
    "messages": [
      {"role": "user", "content": "Review this pull request"}
    ],
    "stream": true
  }'

Estimate token cost

Enter expected input and output tokens to compare current listed rates.

Input tokensOutput tokens

OmniaKey estimate: $0.02475
Official-rate estimate: $0.20
Estimated savings: $0.17525

Estimate covers input and output tokens only. Cache usage, retries, tools, and client-side token accounting can change the final charge.

Related coding guides

Deeper selection, setup, and troubleshooting guidance for this model family.

GPT-6.1 Terra: What's New, Benchmarks and How It Compares to Sol

Half the price of GPT-6.1 Sol, with big gains over GPT-5.6 Terra in computer use, coding and accuracy.

Read guide

GPT-6.1 Terra vs GPT-6.1 Sol: Benchmarks, Speed and Which to Use

Terra for volume and speed, Sol for the hardest coding: the numbers behind the choice.

Read guide

GPT-5.6 Review: Which Model Should You Choose?

GPT-5.6 is a family, not one model. This guide puts Sol, Terra, and Luna's official specifications, prices, and operating boundaries into one practical decision framework.

Read guide

Sources and verification

Model facts and links were checked on 2026-10-27. Provider documentation remains the authority for upstream behavior.

Frequently asked questions

How much does GPT-6.1 Terra cost on OmniaKey?

OmniaKey currently charges $0.1125 USD per 1M input tokens and $0.675 USD per 1M output tokens, in USD. Direct provider rates are separate and may change with input context length.

What is the context window for GPT-6.1 Terra?

The listed context window is 1M. Actual usable limits can also depend on the client and upstream request format.

Is GPT-6.1 Terra better than GPT-6.1 Sol?

Not for the hardest work. OpenAI still names Sol and Astra for the most complex coding, and Sol holds the higher tier. Terra's advantage is volume: it lists lower rates for input, output and cache reads, accepts the none effort setting for tasks that need no reasoning, and carries a knowledge cutoff one month later than Sol's.

How do I turn off reasoning on GPT-6.1 Terra?

Set reasoning effort to none. Terra accepts the full range from none to max, while Sol starts at low, so Terra is the GPT-6.1 model for classification, short rewrites and chat replies where reasoning tokens would only add cost. Check that your client does not send a default of its own, and drop temperature and top_p when reasoning is on.

How is GPT-6.1 Terra billed?

By input, output and cached-input tokens, with every Terra rate listed below the matching GPT-6.1 Sol rate. Reasoning tokens are billed as output, and cached input has its own rate. Batch and Flex are listed below Standard and Fast mode above it. Current rates appear in the price table on the model page.

Can GPT-6.1 Terra read images or edit video?

It reads text and images and returns text. OpenAI's launch post lists video editing among the tasks Terra handles, describing a model that operates video software through computer use rather than one that returns video. Audio and video input are not supported, so frame grabs and subtitles have to be prepared outside the request.

Which clients support gpt-6.1-terra?

Send POST /v1/chat/completions to https://api.omniakey.com/v1 with model set to gpt-6.1-terra, or use POST /v1/responses when the request needs tools or computer use. The catalog lists Codex, Cursor, Cline and Aider as compatible clients, configured with the OmniaKey base URL and key. Copy the id exactly, because it is case-sensitive.