OpenAI's GPT-6.1 Terra targets computer work at a lower price

OpenAI released GPT-6.1 Terra on Oct. 26 as the new middle tier of its model lineup, replacing GPT-5.6 Terra. The company describes it as a fast, efficient model for visual reasoning, computer use and clearer writing, aimed at routine office tasks such as filling out forms, drafting proposals and reading technical drawings.

Standard API pricing is $1 per million input tokens, $0.05 per million cached input tokens and $5 per million output tokens, under the name gpt-6.1-terra. OpenAI says that is 50% below GPT-5.6 Terra and GPT-6.1 Sol. Artificial Analysis lists the previous Terra at $2 and $12, so the output price fell more than the input price.

Most performance figures come from OpenAI. It says Terra beats GPT-5.6 Terra by 15 percentage points on Agents' Last Exam at medium effort, at about a third of the cost, and costs roughly nine times less per task than Claude Opus 5.5 at maximum effort. It also reports that the error rate on its internal set of error-inducing prompts fell from 53.6% to 16.7% at low effort.

Artificial Analysis, which runs its own tests, reported a 6-point gain on its Intelligence Index over GPT-5.6 Terra at maximum effort, at $0.40 per task against $1.40, and a hallucination rate down from 88% to 38%. OpenAI says Sol and GPT-6 Astra remain the better choice for the most complex coding tasks, and TNW noted that the Opus comparison rests on OpenAI's own testing.

Terra is listed for Plus, Pro, Enterprise and Edu users in ChatGPT Work and Codex, while the Chat tab began moving to GPT-6 Sol and GPT-6 Luna on Oct. 7. OpenAI says Terra meets the Critical cybersecurity threshold under its Preparedness Framework and keeps the protections built for GPT-6 Astra. Because the effort setting changes both price and results, cost per finished task on a real workload is the useful comparison.