GPT-6.1 Terra API: Pricing, Specs, and How to Use It in Codex

Short answer

GPT-6.1 Terra (gpt-6.1-terra) is the GPT-6 model OpenAI released on October 26, 2026 for visual reasoning, precise computer use, and clear, natural writing. By price it sits between GPT-6.1 Sol and GPT-6 Luna: pick it when the work is about images, screens or text, and keep Sol for complex coding. The official price is $1 per million input tokens and $5 per million output tokens, half of Sol's list price, with a 1.05M-token context window. On TeamoRouter it uses the same model ID and is billed per token with no subscription (live prices are on the pricing page). It works in Codex CLI 0.159.0 or later by changing the model name in config.toml.

No card at all? You can also top up with USDT, USDC or ETH: click Pay with crypto in the top-up dialog and send it from Binance, OKX or any wallet. The minimum is $5 and no browser extension is needed. Step by step: Pay for Claude and GPT APIs with crypto (USDT).

Specs at a glance

From OpenAI's model page:

OpenAI's examples of what to use it for: matching product photos to catalog entries, interpreting technical drawings, acting on application screens, and drafting and revising text. In the Responses API it supports built-in tools including web search, file search, code interpreter, computer use, apply_patch, MCP and tool search.

Terra, Sol, Luna, or Astra

OpenAI's GPT-6 guide now lists four models. Prices below are OpenAI's standard rates for prompts up to 272K input tokens.

How to choose, based on OpenAI's own guidance:

• Terra when the input is visual or the output is prose: reading screenshots and dashboards, comparing technical drawings, working through an app's screens, drafting proposals or documentation. OpenAI's Codex docs call it "a fast, efficient choice for writing, visual understanding, and computer use."
• Sol for complex coding and multi-step agent work. OpenAI describes it as near-Astra performance at a lower cost.
• Luna for extraction, classification, transformation and structured summaries, when you know what a good result looks like and run many of them.
• Astra for the hardest end-to-end work, when quality matters more than cost.

Terra also supports none reasoning effort, which GPT-6.1 Sol and GPT-6 Astra do not.

GPT-6.1 Terra vs GPT-5.6 Terra. Terra replaces GPT-5.6 Terra and costs less on every line at list price: input $1 vs $2, cached input $0.05 vs $0.20, output $5 vs $12 per million tokens. OpenAI's migration advice is to compare both models on representative images and workflows at the same reasoning effort, and to move any tool calls made through Chat Completions to the Responses API. Codex does not switch models by itself: if your config.toml says gpt-5.6-terra, change it by hand.

Pricing

When a prompt has more than 272K input tokens, the whole request is billed at 2x input and cache rates and 1.5x output. On OpenAI's own API, Batch and Flex are 50% cheaper than standard and Fast mode costs 2x.

TeamoRouter also bills per token, with no subscription. The current numbers are on the pricing page.

Use it in Codex CLI

OpenAI requires Codex CLI 0.159.0 or later for GPT-6.1 Terra. Check with codex --version and update with npm install -g @openai/codex if needed.

Put this in ~/.codex/config.toml and set your TeamoRouter key in OPENAI_API_KEY:

model_provider = "teamorouter"
model = "gpt-6.1-terra"
model_reasoning_effort = "medium"

[model_providers.teamorouter]
name = "TeamoRouter"
base_url = "https://api.teamorouter.com/v1"
env_key = "OPENAI_API_KEY"
wire_api = "responses"

To keep Sol as your default and use Terra only for some sessions, leave model = "gpt-6.1-sol" and start Codex with codex -m gpt-6.1-terra. Inside a session, /model switches models and reasoning effort. OpenAI suggests starting at medium and raising effort only if the better result is worth the extra time and tokens. Step-by-step setup: Codex CLI guide.

Use it in OpenCode

OpenCode connects to TeamoRouter through an OpenAI-compatible provider. Add the provider to opencode.json, put your sk-teamo- key in the TEAMOROUTER_API_KEY environment variable, and list gpt-6.1-terra under models:

{
  "$schema": "https://opencode.ai/config.json",
  "provider": {
    "teamorouter": {
      "npm": "@ai-sdk/openai-compatible",
      "name": "TeamoRouter",
      "options": {
        "baseURL": "https://api.teamorouter.com/v1",
        "apiKey": "{env:TEAMOROUTER_API_KEY}"
      },
      "models": {
        "gpt-6.1-terra": { "name": "GPT 6.1 Terra" },
        "gpt-6.1-sol": { "name": "GPT 6.1 Sol" },
        "gpt-6-luna": { "name": "GPT 6 Luna" }
      }
    }
  }
}

Restart OpenCode and choose GPT 6.1 Terra in the model picker. For OpenCode Desktop, see the OpenCode guide.

The same key and base_url work in other agents with an OpenAI-compatible provider, such as Cline for VS Code, DeepSeek Harness and Cursor. Only the model name changes: set it to gpt-6.1-terra.

Call the API directly

Responses API (use this one if you need tool calling):

curl https://api.teamorouter.com/v1/responses \
  -H "Authorization: Bearer YOUR_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gpt-6.1-terra",
    "reasoning": {"effort": "medium"},
    "input": "Rewrite this release note in plain language for customers: ..."
  }'

OpenAI Python SDK, with an image:

from openai import OpenAI

client = OpenAI(base_url="https://api.teamorouter.com/v1", api_key="YOUR_KEY")
resp = client.responses.create(
    model="gpt-6.1-terra",
    input=[{
        "role": "user",
        "content": [
            {"type": "input_text", "text": "List the dimensions and tolerances shown on this drawing."},
            {"type": "input_image", "image_url": "https://example.com/drawing.png"},
        ],
    }],
)
print(resp.output_text)

For a local file, pass a base64 data URL instead, for example "image_url": f"data:image/png;base64,{b64}".

Chat Completions also accepts gpt-6.1-terra, but OpenAI supports tool calling for this model only in the Responses API. If you reuse older request code, note that with any reasoning effort other than none, temperature, top_p and top_logprobs must be removed. Full reference: API integration.

Paying for usage

No subscription: you top up a balance from $5 and pay only for the tokens you use, by card or with USDT, USDC or ETH from any exchange app or wallet, no browser extension needed. The same key also works for Claude and Gemini.

FAQ

Is GPT-6.1 Terra in ChatGPT? OpenAI makes it available in ChatGPT Work and Codex on Plus, Pro, Business, Enterprise and Edu plans, but not in Chat. Through the API you call it with an API key, and no ChatGPT plan is needed.

Terra or Sol for coding? For complex coding, Sol. OpenAI positions Sol as near-Astra performance for complex work and Terra as the model for writing, visual understanding and computer use. If a task is mostly reading screenshots or writing documentation, that matches what Terra is built for; /model in Codex switches between the two.

Does it support images? Yes, text and image input; output is text only, and audio and video are not supported. OpenAI recommends checking results against the source material, especially names, numbers and details read from images.

How much does it cost? At OpenAI's list price, $1 per million input tokens, $0.05 cached input and $5 output, with 2x input and 1.5x output for the whole request once a prompt passes 272K input tokens. TeamoRouter's current price is on the pricing page.

Which reasoning effort should I use? medium is the default and OpenAI's suggested starting point; raise it only if the better result is worth the added time and tokens. Unlike Sol and Astra, Terra also accepts none.

Codex says the model was not found. Check that Codex CLI is 0.159.0 or later, that config.toml has model_provider = "teamorouter" at the top, and that OPENAI_API_KEY is set in the terminal that launched Codex.

Related

• GPT-6.1 Sol API: the tier above Terra, for complex coding
• GPT-6 Luna API: the cheapest GPT-6 model
• GPT-6 Sol API and Codex setup: config.toml setup for macOS, Windows and Linux
• Codex CLI guide
• OpenCode guide