What Is GPT-6.1 Terra? OpenAI's $1/$5 Computer-Use Model
The short answer
GPT-6.1 Terra is OpenAI’s new middle model: GPT-6-generation quality on everyday and computer-use work at half the price of GPT-6.1 Sol. OpenAI released it on October 26, 2026 as gpt-6.1-terra. It costs $1 per million input tokens and $5 per million output tokens ($0.05 cached input), has a 1.05M-token context window, accepts text and images, and supports reasoning effort from none to max. It replaces GPT-5.6 Terra ($2/$12), which it beats on every benchmark OpenAI published.
Key facts
| GPT-6.1 Terra | |
|---|---|
| Released | October 26, 2026 |
| API model ID | gpt-6.1-terra |
| Price (standard) | $1 input · $0.05 cached input · $1.25 cache write (30 min) · $5 output per 1M tokens |
| Long prompts (>272K input) | 2x input and cache, 1.5x output, whole request ($2 / $7.50) |
| Batch / Flex | 50% off ($0.50 / $2.50) |
| Context | 1,050,000 tokens (922K max input, 128K max output) |
| Modalities | Text and image in, text out (no audio or video) |
| Knowledge cutoff | June 1, 2026 |
| Reasoning effort | none, low, medium (default), high, xhigh, max |
| Where | API (Responses, Chat Completions, Batch); ChatGPT Work and Codex on Plus, Pro, Enterprise, Edu |
Tool calling requires the Responses API. Regional processing adds 10% where available.
What OpenAI says it is good at
OpenAI’s launch post focuses on three areas, all vendor-reported:
- Computer use. On Agents’ Last Exam (professional tasks such as turning architectural drawings into a 3D model), Terra at medium effort scores 15 points above GPT-5.6 Terra at about a third of the cost, and performs on par with GPT-6 Astra at a tenth of the price per token. At max effort, OpenAI says it is about 9x cheaper per task than Claude Opus 5.5, and completes tasks over three times faster than GPT-6 Astra and GPT-6.1 Sol.
- Coding. Terminal-Bench 4.0 at max effort: 50.3%, against 21.5% for GPT-5.6 Terra. For the hardest coding tasks OpenAI still recommends GPT-6.1 Sol or GPT-6 Astra.
- Factuality. On OpenAI’s internal error-inducing prompts, the error rate at low effort fell from 53.6% (GPT-5.6 Terra) to 16.7%; at max effort it is 4.3%.
Artificial Analysis lists GPT-6.1 Terra at 48 on its Intelligence Index — level with GPT-6 Sol, below GPT-6.1 Sol (52), GPT-6 Astra (53) and Claude Opus 5.5 (58).
Where it sits in OpenAI’s lineup
| Model | Input / output per 1M tokens | Use it for |
|---|---|---|
| GPT-6 Astra | $10 / $50 | The hardest reasoning and coding |
| GPT-6.1 Sol | $2 / $10 | Complex coding and agents |
| GPT-6.1 Terra | $1 / $5 | Computer use, documents, images, high-volume agent steps |
| GPT-6 Luna | $0.10 / $0.50 | Classification, extraction, routing |
The nearest rival on price is Claude Sonnet 5.5 at $2/$10 — twice Terra’s price, with stronger coding scores. For full rate cards see our current API prices; for the $2 tier compared head to head, see Gemini 4 Argon vs Claude Sonnet 5.5 vs GPT-6.1 Sol.
Should you switch?
- From GPT-5.6 Terra: yes. It is cheaper on input and output and better on every published benchmark.
- From GPT-6.1 Sol for document, form or browser automation: test it — the price halves and OpenAI’s numbers suggest little quality loss on those tasks.
- For coding agents: keep Sol, Astra or Claude for the planning step and try Terra for cheaper sub-steps.
- Watch the 272K threshold: a long prompt doubles the input rate for the whole request, so keep agent context under it where you can.
OpenAI says Terra meets the “Critical” cybersecurity threshold in its Preparedness Framework and ships with the same safeguards as GPT-6 Astra.
Last verified: October 27, 2026.