GPT-6.1 Terra is OpenAI’s cheaper model aimed at computer-use work: $1 per million input tokens and $5 per million output tokens in US dollars, which OpenAI says is 50% below GPT-5.6 Terra and GPT-6.1 Sol. It is a good fit for high-volume agent tasks, but OpenAI itself still points to other models for the hardest coding. This is an analysis of reported announcements, not a hands-on test.

What was announced
According to BleepingComputer, the model is available to Plus, Pro, Enterprise and Edu users in ChatGPT Work and Codex, and through the API. iThinkDiff reports developers call it as gpt-6.1-terra and that it has been live since October 26, 2026.
Pricing
All prices are US dollars per million tokens for the API.
| Model | Input | Output | Cached input |
|---|---|---|---|
| GPT-6.1 Terra | $1 | $5 | $0.05 |
| GPT-6.1 Sol | $2 | $10 | $0.10 |

Speed and accuracy claims
OpenAI says Terra finishes computer-use tasks more than three times faster than GPT-6 Astra and GPT-6.1 Sol. Reports also cite a drop in error rate at low reasoning effort from 53.6% to 16.7% versus GPT-5.6 Terra. These are vendor-reported numbers; I found no independent benchmarks yet.
Who it matters for
Teams running many repetitive agent tasks (filling forms, operating apps, workplace chores) gain the most, since cost and speed dominate there. For the hardest coding, OpenAI still recommends GPT-6.1 Sol or GPT-6 Astra.
What is still unknown
Independent testing, real-world reliability on your own workflows, and how pricing changes for very large prompts are not yet confirmed. Check OpenAI’s pricing page before budgeting.
For related context, see our look at Google’s Gemini agent and Gemini 4 Argon.

