OpenAI has released GPT-6.1 Terra at half the API price of its previous midrange model. Input costs $1 per million tokens, output costs $5 per million and cached input falls to five cents per million. The model is available through the API and in ChatGPT Work and Codex for Plus, Pro, Enterprise and Education users.
Terra targets office tasks, visual reasoning and computer control. Artificial Analysis places it between GPT-6.1 Sol and GPT-5.6 Terra on its Intelligence Index while reporting a lower estimated cost per task. On that evaluator’s private knowledge test, the model guessed when it lacked an answer 38 percent of the time, down from 88 percent for GPT-5.6 Terra, with a small accuracy increase.
The largest gains appear in computer use and coding. Terra scored 50.3 percent on Terminal-Bench 4.0 at maximum reasoning effort, compared with 21.5 percent for GPT-5.6 Terra. OpenAI nevertheless recommends Sol or Astra for the most demanding coding jobs, reflecting the difference between price efficiency and maximum capability.
OpenAI classifies Terra at its “Critical” cybersecurity threshold and applies safeguards inherited from Astra. Reported prompt-injection defense reached 98.8 percent on OpenAI’s test, which still leaves residual risk and should not substitute for scoped permissions. Buyers should compare complete task cost, latency and error recovery on their own workflows rather than choosing from token price alone.