OpenAI’s new GPT-6.1 Terra model brings API input costs to $1 per million tokens, positioning the system for visual reasoning and computer-based work at half GPT-6.1 Sol’s listed rate.
Terra’s API rates are $1 per million input tokens, $0.05 for cached input and $5 for output. GPT-6.1 Sol lists those services at $2, $0.10 and $10, respectively. For cache writes retained 30 minutes, OpenAI lists a rate of $1.25 per million tokens.
Compared with GPT-5.6 Terra, whose listed rates are $2 for input, $0.20 for cached input and $12 for output, the price cuts vary by token category. OpenAI offers Terra to Plus, Pro, Enterprise and Edu users in ChatGPT Work and Codex, as well as developers through the API under the name gpt-6.1-terra.
The company presents Terra for tasks such as interpreting technical drawings, filling out forms, drafting proposals and editing videos. OpenAI says lower prices could make high-volume automation, including invoice processing, more practical. It also says ChatGPT Work and Codex users can complete more tasks within their plan allowances.
On Agents’ Last Exam, OpenAI reported that Terra at medium reasoning effort scored 15 percentage points higher than GPT-5.6 Terra. The company estimated the comparison cost roughly one-third as much. OpenAI also said Terra matched GPT-6 Astra’s performance on the exam at one-tenth its per-token price. On the tested tasks, Terra took less than a third of the time recorded for GPT-6 Astra and GPT-6.1 Sol.
OpenAI said its internal set of error-inducing prompts produced a factual-error rate of 16.7% for Terra at low reasoning effort, compared with 53.6% for GPT-5.6 Terra. At maximum effort, Terra solved 50.3% of Terminal-Bench 4.0 tasks, versus 21.5% for the older model. The announcement did not provide independent confirmation of these results. OpenAI said GPT-6.1 Sol and GPT-6 Astra remain better suited to its most complex coding work.
