In ChatGPT Work, GPT-6.1 Terra is intended for work that involves interpreting images or screens and taking action in a browser, on a desktop, or in an app. Official examples include matching product photos from a supplier catalog to catalog entries and updating product information, as well as checking campaign assets for cropping and text placement and organizing the required edits.
Suitable tasks: Matching product photos to catalog entries and updating product information; checking campaign assets for cropping and text placement, then organizing the required edits.
Unsuitable tasks: Tasks that require selecting Terra in Chat; the source does not say it will automatically perform specific image edits, and it provides no size or accuracy measurements or guarantees about results.
Applicable model version: GPT-6.1 Terra.
Applicable client, agent, or API: ChatGPT desktop app, ChatGPT Work on the web, and ChatGPT Work on mobile. This article documents a Work client workflow; it is not an API guide.
Recommended reasoning level and parameters: Not specified in the source.
The source does not provide complete prompts that can be copied directly. The checklists below are editorial adaptations of the official examples, not text from OpenAI:
Product catalog matching
Task: Match product photos in the supplier catalog to the existing product catalog, then update the corresponding product information.
Matching criteria: [Enter product features or catalog fields that can be used for verification]
Fields to update: [Enter fields]
Output: List each catalog image, candidate catalog entry, matching rationale, and fields to update. Flag any match that cannot be confirmed.Campaign asset review
Task: Check campaign assets for cropping and text placement, then organize the required edits.
Target placement or dimensions: [Enter the requirements; if they are not provided, first list the information needed]
Checks: Whether the subject is cropped, whether text is obscured, and whether text placement meets the requirements.
Output: List issues and suggested edits for each asset. Flag anything uncertain separately.The following steps are an editorial adaptation of the official examples; the page itself does not provide step-by-step instructions:
In the ChatGPT desktop app or ChatGPT Work on the web, open the model list and select GPT-6.1 Terra under Advanced. ChatGPT Work on mobile also supports Terra.
Before starting, confirm that the required browser or computer-use capabilities have been configured and that the current account has access.
For product matching, specify which supplier catalog and product catalog to use, the available verification fields, and what information to update. Review the matches first, then handle uncertain items.
For campaign assets, provide the target requirements for cropping and text placement. Review each asset for issues and organize the edits into a list.
When finished, verify the catalog updates or edit list, with particular attention to uncertain matches flagged by the model and unclear placement requirements.
OpenAI describes GPT-6.1 Terra as a fast, efficient model for professional work that combines image understanding with precise computer-use capabilities.
The official description covers reasoning about images and screens, and taking action in browsers, on desktops, and in apps.
Official examples include matching product photos from a supplier catalog to catalog entries and updating product information; and checking campaign assets for cropping and text placement, then organizing the edits.
OpenAI says the model is available in ChatGPT Work and Codex on the Plus, Pro, Business, Enterprise, and Edu plans. This article focuses on the Work client.
The model can be selected from the Advanced model list in the ChatGPT desktop app or ChatGPT Work on the web; ChatGPT Work on mobile also supports it.
GPT-6.1 Terra is not available in Chat and does not support UAE reasoning residency.
Enterprise availability depends on workspace availability and administrator controls.
Browser use and computer use each have setup and access requirements; the page does not list the specific configuration steps.
The page does not provide complete prompts, task parameters, accuracy data, or instructions for specific image-editing operations for the examples above. The input frameworks and review steps in this article are editorial adaptations and do not represent an official testing workflow.
This article documents the ChatGPT Work client workflow. Check the OpenAI API models documentation separately for the model API ID, reasoning configuration, and API tool capabilities.
At the time of collection, the “Use GPT-6.1 Terra for visual reasoning and computer use” entry on the OpenAI ChatGPT Learn page for “October 26–30, 2026” was reviewed. The full entry was readable. The prompt frameworks and operating steps were editorial adaptations based on the public examples and were not presented as official source text. The page content may change later.
GPT-6.1 Terra