99.9% SLA
GPT-6.1 Terra: Model Introduction & Practical Guide
GPT-6.1 Terra is OpenAI's balanced model for visual reasoning, coding and writing. It is designed to interpret screenshots, understand technical drawings and support everyday professional work, with a 1,050,000-token context window and up to 128,000 output tokens.
Quick Facts
| Attribute | Value |
|---|---|
| Model | GPT-6.1 Terra |
| Developer | OpenAI |
| Positioning | Balanced visual reasoning, coding and writing |
1. Model Overview
Terra focuses on turning visual understanding into useful analysis and action. OpenAI positions it for recurring work such as reading documents, drafting proposals and working with application screens, with higher-priced models available for its most difficult engineering tasks.
2. Core Features
Visual reasoning. Interpret screenshots, product images and technical drawings.
Writing and coding. Draft clear text, explain implementations and investigate defects.
Reasoning control. Adjust effort from none through max.
Structured workflows. Use schema-constrained output and supported tool interfaces.
3. Technical Specifications
| Specification | Detail |
|---|---|
| Context window | 1,050,000 tokens |
| Maximum output | 128,000 tokens |
| Input | Text and images; OpenRouter also lists file inputs |
| Output | Text |
| Reasoning effort | none / low / medium / high / xhigh / max |
| Default effort | medium |
| Tool calling on OpenAI | Requires the Responses API |
| Native audio or video input | Not supported |
| Knowledge cutoff | June 1, 2026 |
4. Capability Comparison
Within the GPT-6.1 family, Terra emphasizes efficient everyday work and costs less than Sol at standard API rates. OpenAI positions Sol and Astra for its most complex coding work. Choose between them using the task's difficulty and the quality of the resulting deliverable.
5. Core Advantages
A large context window, detailed visual input and adjustable reasoning support a workflow that alternates between short writing tasks and document or screenshot analysis. Structured responses can make the results easier to validate and reuse.
6. Recommended Use Cases
Screenshot interpretation, product-image matching, technical drawing explanations, proposal drafting, code assistance and structured document extraction.
7. Example Prompts
Visual work: "Inspect this screenshot. Describe the visible state, identify the relevant control and state what cannot be determined from the image."
Writing: "Rewrite this proposal for a technical buyer. Preserve the factual claims, shorten repetition and list statements that need supporting evidence."
8. Selection Recommendations
Evaluate the model on representative documents and screenshots. Adjust effort using measured output quality and response time. For agent workflows, verify that the selected endpoint exposes the tools you need; model capability alone does not supply an execution environment.
Sources & Further Reading
Specifications checked on November 5, 2026.
