On this page · 5 sections ▾
OpenAI announced GPT-6.1 Terra on October 26, 2026, a mid-range model aimed at routine computer work such as filling forms, entering invoice data and editing video. It replaces GPT-5.6 Terra, costs 50% less through the API, and is available in ChatGPT Work, Codex and the API, according to The Next Web and Digital Trends.
Key facts
- Pricing: $1 per million input tokens, $5 per million output tokens and $0.05 per million cached input tokens, per FoneArena. Digital Trends lists GPT-6 Astra at $10 and $50.
- Availability: ChatGPT Plus, Pro, Enterprise and Edu users in ChatGPT Work and Codex, plus the API under the identifier gpt-6.1-terra.
- Other channels: India Today reports rollout through GitHub Copilot and Amazon Bedrock.
- Lineup position: Terra sits above Luna and below Sol and Astra.
- Speed: OpenAI says it completes computer tasks more than three times faster than GPT-6 Astra and GPT-6.1 Sol.
What changed
The headline feature is computer use. Terra reads what is on a screen and then clicks, types and moves between applications to finish multi-step tasks. OpenAI's examples include copying invoice details into accounting software, drafting proposals, interpreting technical drawings and editing videos.
On Agents' Last Exam, a benchmark of professional computer tasks, OpenAI says Terra scored 15 percentage points above GPT-5.6 Terra at medium reasoning effort, at about one-third of the cost per task. FoneArena reports it matched GPT-6 Astra on that test at one-tenth of the price per token.
On the company's internal set of error-inducing prompts, the error rate at low reasoning effort fell from 53.6% to 16.7%. At maximum effort it fell 18 percentage points to 4.3%, close to Astra and Sol, according to 9to5Mac and FoneArena.
In coding, Terra solved 50.3% of tasks on Terminal-Bench 4.0 at maximum effort, against 21.5% for its predecessor. OpenAI still points users to Sol or Astra for the most complex coding work.
How it compares
FoneArena lists OpenAI's comparison with Anthropic's Claude Opus 5.5 at maximum reasoning effort: more than three times the token efficiency, one-fourth the price per token and roughly nine times lower cost per task. The Next Web notes these results come from OpenAI's own testing and have not been independently checked. The same caution applies to the error-rate and benchmark figures above.
Safety and context
OpenAI says Terra reaches the Critical level for cybersecurity under its Preparedness Framework and carries the same safeguards as GPT-6 Astra. It also says the model is more likely than GPT-5.6 Terra to stay within the scope of a task and to disclose its limits, and that it resisted malicious instructions in prompt-injection tests.
The Next Web adds that the release follows OpenAI's cancellation last month of GPT-6.1 Astra after it failed internal safety tests. The other outlets do not mention this.
