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

OpenAI announced GPT-6.1 Terra, which it describes as a fast, efficient model that combines detailed visual reasoning and precise computer use with clearer, more natural writing. OpenAI calls it a significant step up from GPT-5.6 Terra and says it is aimed at the communication and computer-based tasks of the workday, from drafting proposals and interpreting technical drawings to filling out forms and editing videos.

The pitch is cost as much as capability. OpenAI says Terra's standard API prices are 50% lower than those of GPT-5.6 Terra and GPT-6.1 Sol, which it argues makes high-volume automation practical, such as reading incoming invoices and entering the details into accounting software. In ChatGPT Work and Codex, the same usage allowance covers more tasks than with GPT-5.6 Terra, and substantially more than with GPT-6.1 Sol or GPT-6 Astra.

On computer use, OpenAI cites Agents' Last Exam, a benchmark of professional tasks that includes turning architectural drawings into a 3D model and reviewing an animated film shot by shot to produce a time-coded log for a restoration team. At medium reasoning effort, Terra 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 its price per token. At max effort, OpenAI says Terra is over three times as token efficient as Claude Opus 5.5 and one-fourth its price per token, giving a ~9x lower cost per task. It also completes tasks over three times faster than both GPT-6 Astra and GPT-6.1 Sol.

On MMMU-Pro, a visual reasoning benchmark, Terra beats GPT-5.6 Terra at every reasoning setting and nearly matches GPT-6 Astra, OpenAI's most capable frontier model, and GPT-6.1 Sol at much lower cost. On OSWorld 2.1, which tests how well agents use desktop apps for multistep tasks, Terra at medium effort surpasses GPT-5.6 Terra's best score at a fraction of the cost per task.

OpenAI also reports gains in factuality. On its 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, a drop of nearly 37 percentage points. At maximum effort the rate fell by 18 percentage points to 4.3%, nearly matching GPT-6 Astra and GPT-6.1 Sol.

For software engineering, Terminal-Bench 4.0 tests coding agents on command-line problems such as fixing a webpage whose layout shifts as it loads. At maximum reasoning effort, Terra solves 50.3% of tasks, more than double GPT-5.6 Terra's published score of 21.5%. OpenAI adds that for the most complex coding tasks, GPT-6.1 Sol and GPT-6 Astra remain the better choice. On Terminal-Bench-Science 0.1, which covers scientific research tasks such as tracking cell divisions in a time-lapse video, Terra at max reasoning surpasses GPT-6.1 Sol at lower cost and comes within five points of Opus 5.5 at less than one-fifth the estimated cost.

On safety, OpenAI says GPT-6.1 Terra meets the Critical threshold in cybersecurity under its Preparedness Framework and keeps the protections developed for GPT-6 Astra. It carries the alignment improvements of the GPT-6 series: per OpenAI's evaluations it is more likely than GPT-5.6 Terra to stay within scope and honestly disclose limitations and potential blockers. In OpenAI's prompt-injection tests it showed strong resistance to malicious instructions meant to divert it from the user's task. A system card has more detail.

Availability: starting the day of the announcement, 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. Standard API pricing is $1 per million input tokens, $0.05 per million cached input tokens and $5 per million output tokens.

Key facts

GPT-6.1 Terra is available in ChatGPT Work and Codex to Plus, Pro, Enterprise and Edu users, and in the API as gpt-6.1-terra.
Standard 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 GPT-5.6 Terra and GPT-6.1 Sol.
On Agents' Last Exam at medium effort, OpenAI says Terra scores 15 percentage points higher than GPT-5.6 Terra at about one-third the cost, and matches GPT-6 Astra at one-tenth the price per token.
On OpenAI's internal error-inducing-prompt benchmark, the low-effort error rate fell from 53.6% to 16.7%; on Terminal-Bench 4.0 at maximum effort Terra solves 50.3% of tasks versus 21.5% for GPT-5.6 Terra.
For the most complex coding tasks, OpenAI says GPT-6.1 Sol and GPT-6 Astra remain the better choice.

Why it matters

OpenAI frames Terra as a move along the cost-intelligence frontier: a new balance of speed, capability and cost. The headline claims are that it beats GPT-5.6 Terra on computer use and coding while costing half as much per token, and that on Agents' Last Exam it reaches GPT-6 Astra's level at medium effort for one-tenth the per-token price. Against Claude Opus 5.5 at max effort, OpenAI claims a ~9x lower cost per task. If those hold up, a cheaper tier does work that previously needed OpenAI's larger models.

Who it affects

ChatGPT Work and Codex users on Plus, Pro, Enterprise and Edu plans get the model at once, and OpenAI says their usage allowance stretches further than with GPT-5.6 Terra, GPT-6.1 Sol or GPT-6 Astra. Developers building high-volume automation, such as reading invoices and entering them into accounting software, are the group OpenAI names for the lower API prices. Teams that compare vendors on cost per task will look at the Claude Opus 5.5 comparison.

How to use it

In ChatGPT Work and Codex, pick GPT-6.1 Terra if your plan is Plus, Pro, Enterprise or Edu. In the API, use the model name gpt-6.1-terra. Standard prices are $1 per million input tokens, $0.05 per million cached input tokens and $5 per million output tokens. OpenAI's own guidance is to use Terra for fast computer use, visual reasoning, document and communication work, and to stay with GPT-6.1 Sol or GPT-6 Astra for the most complex coding tasks. A system card is linked from the announcement.

How solid is it

This is a first-party launch post from OpenAI, so every benchmark is OpenAI's own claim. The error-rate figures come from OpenAI's own internal benchmark, not an independent one. No absolute benchmark scores are given for Agents' Last Exam, MMMU-Pro or OSWorld 2.1, only relative comparisons. No per-token prices are stated for GPT-6.1 Sol, GPT-6 Astra, GPT-5.6 Terra or Claude Opus 5.5, so the price ratios cannot be checked from the post. No independent third-party evaluation or outside reaction is reported.

Risks and caveats

OpenAI says GPT-6.1 Terra meets the Critical threshold in cybersecurity under its Preparedness Framework, and relies on safeguards carried over from GPT-6 Astra. Even at maximum effort the internal error rate is 4.3%, and at low effort 16.7%. OpenAI concedes Terra is not the best option for the most complex coding work. No context window size or parameter count is stated, and the release gives no calendar date beyond 'starting today'.

“Terra continues to advance the cost-intelligence frontier across our model families, providing a new balance of speed, capability, and cost.”

— OpenAI, GPT-6.1 Terra announcement