OpenAI launches GPT-6.1 Terra at $1/$5 per million tokens

GPT-6.1 Terra is OpenAI's new workday model: detailed visual reasoning and precise computer use, plus clearer, more natural writing. OpenAI positions it as a significant step up from GPT-5.6 Terra, aimed at the communication and computer-based tasks that fill a workday rather than at the hardest coding or research jobs, where it still points users to GPT-6.1 Sol or GPT-6 Astra.

The pitch is practical, not abstract: a model that can look at a technical drawing, a form, or a video timeline and then do something with it. OpenAI's own example of high-volume work is reading incoming invoices and entering their details into accounting software. OpenAI presents that as the kind of high-volume work developers can now automate at a lower price.

Standard API pricing is $1 per million input tokens, $0.05 per million cached input tokens, and $5 per million output tokens. OpenAI says that is 50% lower than the standard prices for both GPT-5.6 Terra and GPT-6.1 Sol. On OpenAI's pricing page (read October 26, 2026), GPT-5.6 Terra is $2 input, $12 output, and $0.20 cached input per million tokens, and GPT-6.1 Sol is $2 and $10, so the input price is exactly half of both, the output price is exactly half of Sol's, and it is about 58% below GPT-5.6 Terra's. Separately, OpenAI says the same usage allowance in ChatGPT Work and Codex covers more tasks with Terra than with GPT-5.6 Terra, and substantially more than with GPT-6.1 Sol or GPT-6 Astra. Prompts over 272K input tokens are priced higher, and Batch and Flex prices are 50% below Standard, per OpenAI's model page.

Starting October 26, GPT-6.1 Terra is available to all Plus, Pro, Enterprise, and Edu users in ChatGPT Work and Codex. Developers can call it through the OpenAI API as gpt-6.1-terra. OpenAI's announcement names only those four plans; it does not mention Free, Go, Business, or Team, so do not assume access on those plans.

On Agents' Last Exam, a benchmark of professional computer-use tasks, such as turning architectural drawings into a 3D model and building a time-coded log from an animated film, GPT-6.1 Terra at medium reasoning effort scores 15 percentage points higher than GPT-5.6 Terra at about one-third the cost, and performs on par with GPT-6 Astra at one-tenth Astra's price per token. At maximum effort, OpenAI says Terra is over three times as token-efficient as Claude Opus 5.5 and one-fourth its price per token, for roughly 9x lower cost per task, and finishes tasks over three times faster than both Astra and GPT-6.1 Sol. These are OpenAI's own measurements, on its own benchmark, at the effort settings it chose.

The factuality and coding numbers tell a similar story of a cheaper model catching up. On OpenAI's internal benchmark of error-inducing prompts, the error rate at low reasoning effort fell from 53.6% with GPT-5.6 Terra to 16.7% with GPT-6.1 Terra, and at maximum effort fell by 18 percentage points to 4.3%. On Terminal-Bench 4.0, Terra at maximum effort solves 50.3% of command-line coding tasks, more than double GPT-5.6 Terra's published 21.5%. OpenAI is explicit that Sol and Astra remain the better choice for the most complex coding tasks.

Artificial Analysis, which runs its own evaluations, reported results on October 26. It says that at maximum effort, GPT-6.1 Terra gains 6 points over GPT-5.6 Terra on its Intelligence Index at less than a third of the cost per task ($0.40 versus $1.40), scoring 4 points below GPT-6.1 Sol and 10 points above GPT-6 Luna. On its Coding Agent Index v2.0, Terra scores 48, versus 54 for Sol and 31 for Luna. On Artificial Analysis' own hallucination test, Terra guesses 38% of the time when it does not know an answer, down from 88% for GPT-5.6 Terra, while accuracy rose 2 points. Artificial Analysis also reports a gain of roughly 20 points on Terminal-Bench 4.0 and says gains on most other evaluations are smaller, for example 5 points on GDP.pdf. Treat both sets of numbers as evidence to try the model on your own tasks, not as proof it will beat a more expensive model on yours.

According to OpenAI, GPT-6.1 Terra “meets the Critical threshold in cybersecurity” under its Preparedness Framework and maintains the protections developed for GPT-6 Astra. OpenAI’s system card addendum says it is “treating GPT-6.1 Terra as Critical capability in Cybersecurity” and High capability in the Biological and Chemical domain, applies the same safeguards stack as GPT-6.1 Sol and GPT-6 Astra, and defines Critical capability as a model that can “identify and develop functional zero-day exploits of all severity levels in many hardened real-world critical systems without human intervention” or devise end-to-end novel cyberattack strategies from a high-level goal. It is OpenAI’s own determination, not an independent finding. The company also reports stronger resistance to prompt injection and a greater likelihood that the model stays within scope and honestly discloses limitations and blockers, compared with GPT-5.6 Terra. More detail is in the model system card.

The rating is relevant because Terra is built to take action across apps, not only to chat. As editorial advice, not OpenAI guidance: if you connect any agent to real accounts, starting with supervised tasks and reviewing its work is a sensible default.