GPT-6.1 TerraTranslation site

What is GPT-6.1 Terra?

GPT-6.1 Terra is a model released by OpenAI on October 26, 2026, with a focus on visual reasoning, computer operations, and natural writing. It is part of the GPT series but is not simply a rebranded version of GPT-5: when integrating new projects, users should select Terra, Sol, or Astra based on actual task requirements and avoid mixing older parameter versions with newer models. Terra is suitable for everyday tasks that require a balance of speed and cost, while more complex reasoning and large-scale software engineering projects may still benefit from higher-end models.

Key Features and Use Cases

• Understanding Text and Images:Read product photos, charts, and technical drawings, and extract information based on written requirements; suitable for organizing product documentation, verifying invoices, and analyzing design documents.
• Tool Collaboration:By integrating search, file, code, and computer operation tools, information analysis can be incorporated into actual workflows; the model itself does not automatically obtain website accounts, software permissions, or business data.
• Writing and Programming:Drafting proposals, polishing long-form documents, generating structured results, and assisting with code revisions. The output still requires manual verification and testing; official benchmark scores should not be equated with success rates in every business scenario.

How to use

1. 个人用户可在支持该型号的ChatGPT Work或Codex套餐中选择Terra，上传有权使用的资料，说明目标、输出格式和验收标准。
2. Developers set up their projects and API keys on the OpenAI platform, then selectgpt-6.1-terra. Use the Responses API when you need to call the tool, and store the secret key on the server.
3. Start by testing with a small number of real-world tasks—for example, generating forms for review based on images of invoices. After verifying the fields, amounts, and sources, gradually integrate the system into the business operations; establish clear confirmation points for operations such as sending, paying, and deleting.

Pricing and Usage Restrictions

The standard API pricing is $1 per million input tokens and $5 per million output tokens; this base rate applies to requests with inputs of up to 272K tokens. Longer inputs, cached writes, fast mode, and tools may be subject to different rates; you should budget based on actual requests. ChatGPT subscriptions are billed separately from the API. Terra is available to Plus, Pro, Enterprise, and Edu users on Work and Codex, with specific quotas depending on the plan and account.

The model supports text and image input and text output; the tool-assisted video processing should not be interpreted as a native audio-video model. The context size is 1,050,000 tokens, with a maximum output of 128,000 tokens; tool invocation requires the Responses API, and fine-tuning is not supported. It is not an open-weight model available for download and local deployment; service availability is subject to supported regions, account eligibility, and rate limits. Online retrieval, human review, and code testing remain necessary.

How to Choose Among Similar Models

When making a selection, compare the quality of task completion, the amount of manual rework, and the total cost, rather than just comparing a single ranking or the price per token.

Official Information

• Introduction to the Release of GPT-6.1 Terra,API Release Notes (October 26, 2026)
• Model Specifications and Prices,GPT Series Selection Guide
• Gemini 4 Argon Documentation,Introduction to Claude Sonnet 5.5