OpenAI has released GPT-6.1 Terra, a fast, lower-cost model built for computer use, visual reasoning and clear writing. The headline claims are aggressive: on OpenAI's own benchmark it performs on par with the flagship GPT-6 Astra at one-tenth the price per token. Here is what OpenAI says, and what to check before you switch.
Price and availability
Terra launched on 26 October 2026. API pricing is $1 per million input tokens, $0.05 per million cached input tokens and $5 per million output tokens, which OpenAI says is 50% lower than both GPT-5.6 Terra and GPT-6.1 Sol. Batch and Flex processing are 50% below standard, while Fast mode costs twice the standard rate. Prompts above 272,000 input tokens cost more. The model has a context window of about 1.05 million tokens, according to developer coverage.
It is available now to Plus, Pro, Enterprise and Edu users in ChatGPT Work and Codex, and to developers through the API as gpt-6.1-terra.
OpenAI's performance claims
- At medium reasoning effort, Terra scores 15 percentage points higher than GPT-5.6 Terra at about one-third of the cost.
- It matches GPT-6 Astra on the Agents' Last Exam benchmark at one-tenth the price per token.
- At maximum effort, OpenAI says it is over three times as token-efficient as Anthropic's Claude Opus 5.5 at one-fourth the price per token, about nine times lower cost per task, and comes within five points of Opus 5.5 at less than one-fifth of the estimated cost.
- It completes tasks over three times faster than GPT-6 Astra and GPT-6.1 Sol.
These are OpenAI's own internal tests. Independent evaluations of Terra had not yet been widely published at the time of writing, so treat the comparisons with rival models as a vendor claim.
A reality check from early testing
Automation platform Activepieces ran Terra on three simple everyday tasks: sorting tickets, extracting invoice fields and summarising emails. It found a median response time of 1.2 seconds against 0.8 seconds for the older, smaller GPT-5.4 mini, and a cost per task of about $0.000256 against $0.000141. In other words, for very simple, high-volume jobs, a smaller and cheaper model can still win on both speed and cost.
Who it is for
OpenAI pitches Terra at high-volume repetitive work, such as reading incoming invoices and entering their details into accounting software, where a model that is good enough and cheap beats one that is slightly better and expensive. Its computer-use ability, which lets it operate browser and desktop apps, is the main new feature.
Why it matters
Terra lands in the middle of a price war between AI labs, and it makes agent-style automation cheaper for small teams. If you run an app with real traffic, the practical advice is simple: test Terra on your own tasks, measure quality, speed and cost per task, and compare it with your current model rather than relying on launch charts.
Sources: OpenAI, "Introducing GPT-6.1 Terra" (26 October 2026); DataCamp; TrueFoundry; Activepieces.
