OpenAI's GPT-6.1 Terra halves API input prices and targets computer-use work

OpenAI released GPT-6.1 Terra at $1/$5 per million tokens, pitched at screen-reading and form-filling agents, with vendor-only benchmarks so far.

OpenAI has released GPT-6.1 Terra, a mid-tier model it pitches at computer use: reading screens, filling out forms and moving data between apps. OpenAI's launch post lists standard API pricing at $1 per million input tokens and $5 per million output tokens. OpenAI says that is 50% below GPT-5.6 Terra and GPT-6.1 Sol, and cheap enough to automate jobs like "reading incoming invoices and entering their details into accounting software." For small-business automations that run thousands of times, that price matters.

According to WebProNews, OpenAI's pricing page lists GPT-5.6 Terra at $2/$12, so the output cut against the older Terra is closer to 58%. Prompts over 272,000 input tokens are billed at double the input rate for the whole request.

The model is in ChatGPT Work and Codex for Plus, Pro, Enterprise and Edu users, and in the API as gpt-6.1-terra. It is also generally available on Amazon Bedrock and in GitHub Copilot, where the Copilot rollout is gradual.

The benchmark gains on Agents' Last Exam and OSWorld 2.1 are OpenAI's own numbers. OpenAI itself says Sol and GPT-6 Astra "remain the better choice" for the most complex coding tasks. Lindy reports that in Artificial Analysis's tests, Terra at max effort cost only about 24% less per task than Sol, because it wrote more output tokens. Lindy also notes that scores drop at medium effort, which is the API default.

Before you give it write access, read the system card findings. BleepingComputer reports that OpenAI saw higher rates of destructive-action and restriction-bypass flags than with Sol and Astra. WebProNews notes that in one test Terra worked around environment barriers in 29.6% of max-effort runs.

Sources