GPT-6.1 Terra: OpenAI's fast, cheap model for everyday work
GPT-6.1 Terra: OpenAI's fast, cheap model for everyday work
OpenAI's new mid-tier model, released 26 October 2026, is aimed at the computer-based tasks that fill a workday: forms, drawings, documents, video. It costs half as much per token as GPT-5.6 Terra and GPT-6.1 Sol, and OpenAI says it matches GPT-6 Astra on its computer-use benchmark at a tenth of the price.
Links
- Announcement: https://openai.com/index/introducing-gpt-6-1-terra/ (26 Oct 2026)
- System card addendum: https://deploymentsafety.openai.com/gpt-6-1-terra
Headline figures
| Measure | GPT-6.1 Terra |
|---|---|
| API price | $1 / $5 per million input / output tokens |
| Agents' Last Exam (computer use), medium effort | +15 points over GPT-5.6 Terra at about a third of the cost |
| Terminal-Bench 4.0 (coding), max effort | 50.3% (GPT-5.6 Terra 21.5%) |
| Error rate on hard factual prompts, low effort | 16.7% (GPT-5.6 Terra 53.6%) |
Source: OpenAI (vendor figures).
OpenAI says it is about nine times cheaper per task than Claude Opus 5.5 at max effort, and over three times faster than Astra and Sol. For the hardest coding work it still recommends Sol or Astra. It meets the "Critical" cybersecurity threshold in OpenAI's Preparedness Framework and keeps Astra's protections.
Available in ChatGPT Work and Codex for Plus, Pro, Enterprise and Edu, and in the API as gpt-6.1-terra.