OpenAI released GPT-6.1 Terra on the API on 26 October 2026[1]. OpenAI describes it as a fast, efficient model for visual reasoning, precise computer use and clear, natural writing, and it is available as gpt-6.1-terra through the Responses API and the Chat Completions API[1,2].
Per million tokens, it costs $1 for input, $0.05 for cached input and $5 for output, exactly half of GPT-6.1 Sol, which came out on 29 September at $2 input, $0.10 cached input and $10 output[1,2,3]. That is one tenth of the top model, GPT-6 Astra, at $10 input and $50 output, and ten times the cheapest, GPT-6 Luna, at $0.10 input and $0.50 output[3].
Cache writes cost $1.25 with 30-minute retention and $2.50 with 12-hour retention[2,3]. Requests with more than 272K input tokens pay twice the input and cache rates and 1.5 times the output rate for the whole request, Batch and Flex cost half the standard price, fast mode costs double, and regional processing adds 10%[2].
Inputs can be text and images, output is text only, the context window is 1,050,000 tokens and a single response can run up to 128,000 tokens[2]. The knowledge cutoff is 1 June 2026, and reasoning effort can be set to none, low, medium, high, xhigh or max, with medium as the default[2]. Tool calling requires the Responses API, and the model is not available on the free usage tier[2].
In benchmarks published by OpenAI, GPT-6.1 Terra at medium effort scored 15 points higher than GPT-5.6 Terra, which became generally available in July, on Agents' Last Exam, a test of computer use in professional work, at about a third of the cost[4]. At maximum effort it scored 50.3% on Terminal-Bench 4.0, more than double GPT-5.6 Terra's 21.5%, and on an internal test of prompts designed to induce mistakes, the error rate at low effort fell from 53.6% to 16.7%[4].
At maximum effort, OpenAI says the model was more than three times as token-efficient as Anthropic's Claude Opus 5.5 at about a ninth of the cost per task, and finished tasks more than three times faster than GPT-6 Astra and GPT-6.1 Sol[4]. OpenAI still recommends GPT-6.1 Sol and GPT-6 Astra for the most complex coding tasks[4].
On Hacker News, OpenAI's Ted Sanders wrote that the company's models are ordered by size as Luna, Terra, Sol and Astra, and recommended as a starting point picking Astra if you care only about quality, Sol if you are a bit price sensitive, and Terra or Luna if you are very price sensitive[5]. He said GPT-6.1 Terra is probably better than Astra and Sol at technical communication, similar on vision and computer use but much faster, and a bit worse on most other things, which makes it a midrange option for API developers who want something cheaper than Sol but a lot smarter than Luna[5].
The developer Simon Willison compared Terra's place to that of Sonnet in Anthropic's lineup, noting that being neither the cheapest nor the most expensive makes it harder to decide when to use it[6]. His pelican-on-a-bicycle drawing test at max effort cost 11.3161 cents and took 4 minutes and 11 seconds[6].
In ChatGPT, the model is available in Work and Codex for Plus, Pro, Enterprise and Edu users[4]. On safety, OpenAI judged that it reached the Critical threshold for cybersecurity under its Preparedness Framework, and it carries over the safeguards built for GPT-6 Astra[4].
