GPT-6.1 Terra Is Live: Prices, Visual Workflows and API Access

GPT-6.1 Terra release date and availability

The October 26 release replaces the earlier leak-based status of this article. OpenAI now publishes a model page, pricing and API access. A leaked evaluation configuration is no longer needed to establish whether Terra exists.

OpenAI also lists Terra in ChatGPT Work and Codex for Plus, Pro, Business, Enterprise and Edu. Choose it from Advanced where available. The launch update says it is not available in ordinary Chat; workspace controls and tool setup still apply. Codex CLI requires version 0.159.0 or later.

APIMaster's route is listed for API use today. A marketplace listing does not establish compatibility with every OpenAI-hosted tool; check the selected route before moving an existing agent workflow.

GPT-6.1 Terra specifications

The official model documentation lists streaming, Structured Outputs, image input and tool support. Tool calling requires the Responses API, even though plain Chat Completions is supported. Image-generation tool support does not mean Terra natively outputs images.

GPT-6.1 Terra API pricing

Official rates below are USD per million tokens. Short context means at most 272K input tokens. Above that threshold, long-context rates apply to the entire request.

Standard short-context cache writes cost $1.25 for 30-minute retention or $2.50 for 12-hour retention, per million tokens. These are separate from cache reads. See OpenAI pricing.

Compared with GPT-5.6 Terra's Standard short-context $2 / $12, Terra 6.1's input rate is 50% lower and output rate about 58.3% lower. Both rates are half GPT-6.1 Sol's official $2 / $10. These are token-price comparisons, not measured task-cost or quality comparisons.

APIMaster's Terra route currently matches $1 / $5. Discounted Sol routes can have a lower listed price than this Terra route, despite Sol's higher official price. Compare live cards rather than assuming the official tier order determines your bill.

Three visual workflows to try

OpenAI names supplier product photos, technical drawings, application screens and campaign assets as use cases. The prompts below extend those descriptions; they are reproducible workflow suggestions, not published benchmark results or our own completed tests.

Match a supplier lookbook to existing SKUs

Provide a supplier's product photos, your catalog reference images and a SKU specification table. Ask Terra to separate exact matches from similar-looking variants.

Match each supplier photo to a catalog SKU. List visible supporting features, conflicting features and an outcome: confirmed, needs another photo, or no match. Do not merge products with different connectors or accessories. Do not guess unreadable model numbers. Finish with a list of photos the supplier should provide.

The useful result is a matching table with evidence and unresolved cases, rather than an unsupported claim that everything matches.

Review campaign assets across aspect ratios

Provide landscape, square and portrait creatives, the promotion rules and the placement's safe-area requirements.

Check each creative against the promotion rules. Verify visible prices, discount conditions, dates, logo placement, cropping and text overlap. Name the file and region for each issue, quote the evidence and propose a specific edit. Mark unreadable text instead of inventing it.

This turns visual inspection into an actionable revision list. OpenAI explicitly mentions cropping and text placement; price-rule checking is our suggested extension.

Check a multilingual landing page before launch

Provide desktop and mobile screenshots for several languages, plus the approved pricing and CTA destinations.

Compare these screenshots for truncated prices, broken buttons, text overflow and obstructed product images. Identify the screenshot and affected area for every issue. Distinguish valid localization differences from errors. Mark link destinations as unverified unless browser tools actually open them.

With a configured browser or computer-use tool, the workflow can continue into a staging site to verify a suspected link problem. Image input alone does not grant control of an application.

Is GPT-6.1 Terra better than GPT-5.6 Terra?

The confirmed changes include lower official token rates and a newer knowledge cutoff. The 1.05M context window is not new: GPT-5.6 Terra already has that capacity.

The official pages reviewed for this update do not provide a quantified Terra-versus-Terra benchmark or speed multiplier. OpenAI's migration guidance recommends comparing representative images and workflows at the same reasoning effort. Use the same inputs and scoring rules, and record misses, false positives, latency and cost before claiming a quality advantage.

Call GPT-6.1 Terra through APIMaster

Register, create an API key, and set the API base to https://apimaster.ai/v1. For a basic text request:

curl "https://apimaster.ai/v1/chat/completions" \
  -H "Authorization: Bearer $APIMASTER_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gpt-6.1-terra",
    "messages": [
      {"role": "user", "content": "Draft a concise checklist for reviewing multilingual landing-page screenshots."}
    ],
    "reasoning_effort": "medium"
  }'

This is a basic integration example, not a recorded test response. For tools, use Responses and confirm route compatibility; a model-name change alone does not guarantee an existing tool workflow will work unchanged.

FAQ

Has GPT-6.1 Terra been released?

Yes. OpenAI's API changelog records the release on October 26, 2026. The API model ID is gpt-6.1-terra.

Can I use GPT-6.1 Terra on APIMaster now?

Yes. On October 27, APIMaster lists one active public route at $1 input / $5 output per million tokens. Check the current model card.

Is there an APIMaster discount for Terra?

Not at this check: the Terra route matches the official Standard short-context rate. Other GPT routes in the comparison table have their own discounts.

Does Terra work with Chat Completions tools?

OpenAI supports Chat Completions for this model, but tool calling requires Responses. Check the selected provider route's supported features as well.

Does a bigger context window prove an upgrade over GPT-5.6 Terra?

No. Both list 1,050,000 tokens. A quality comparison needs matched tasks and settings.

Updated October 27, 2026. Official specifications and marketplace prices can change. The visual prompts above are suggested workflows, not performance measurements.