OpenAI released GPT-6.1 Terra on October 26, a model for computer use, visual reasoning, coding, and writing, priced at $1 per million input tokens and $5 per million output tokens in its API, according to OpenAI's announcement. The model's API name is gpt-6.1-terra.
OpenAI said GPT-6.1 Terra's standard API prices are "50% lower than those of GPT-5.6 Terra and GPT-6.1 Sol." Against GPT-6.1 Sol, that holds for every rate. Against GPT-5.6 Terra, the previous Terra model, the cut is deeper on output. OpenAI's API pricing page lists GPT-5.6 Terra at $2 per million input tokens and $12 per million output tokens, so the new model's output price is 58% lower. Cached input falls from $0.20 to $0.05 per million tokens, a 75% cut.
Terra is also available to Plus, Pro, Enterprise, and Edu users in ChatGPT Work and Codex, OpenAI said. The model takes text and images and returns text, with a context window of 1,050,000 tokens, up to 128,000 output tokens per response, and a knowledge cutoff of June 1, according to its model page.
What GPT-6.1 Terra costs against OpenAI's other models
The table prices one hypothetical month on each model: 1 billion input tokens and 100 million output tokens at standard rates, with no prompt caching and no prompt over 272,000 tokens. The volumes are an illustration; the prices are OpenAI's.
Model | Input / output ($ per 1M tokens) | Sample month ($) | Cost vs. GPT-6.1 Terra |
|---|---|---|---|
GPT-6 Astra | 10 / 50 | 15,000 | 10x |
GPT-5.6 Terra | 2 / 12 | 3,200 | 2.1x |
GPT-6.1 Sol | 2 / 10 | 3,000 | 2x |
GPT-6.1 Terra | 1 / 5 | 1,500 | 1x |
GPT-6 Luna | 0.10 / 0.50 | 150 | 0.1x |
Source: SaasBuzz calculation from OpenAI's API pricing page (standard tier, prompts of 272,000 tokens or fewer, no caching), checked October 29.
On that mix, moving from GPT-5.6 Terra to GPT-6.1 Terra cuts the bill by 53%. Without caching, the saving runs from 50% for work that is nearly all input to 58% for work that is nearly all output, and output includes reasoning: the tokens a model spends working through a problem before it answers are billed as output tokens, according to OpenAI's reasoning guide. The table holds token counts equal; in practice, models can spend different amounts of reasoning on the same job, which changes the bill.
Other rates follow fixed rules on the model page. Prompts over 272,000 input tokens are billed at twice the input rate and 1.5 times the output rate, or $2 and $7.50 per million tokens for Terra. Batch and Flex processing cost half the standard rate, $0.50 and $2.50, and Fast mode costs double, $2 and $10.
What OpenAI says changed
OpenAI said Terra scored 15 percentage points higher than GPT-5.6 Terra on Agents' Last Exam, a benchmark of professional computer tasks, at medium reasoning effort and about one-third the cost. On its internal set of prompts built to provoke errors, Terra's error rate at low effort was 16.7%, down from 53.6% for GPT-5.6 Terra. On Terminal-Bench 4.0, a test of coding from the command line, Terra solved 50.3% of tasks at maximum effort, against 21.5% for GPT-5.6 Terra. The results are OpenAI's own; SaasBuzz has not tested the model.
The announcement also says where Terra is not the first pick.
For the most complex coding tasks, GPT-6.1 Sol and GPT-6 Astra remain the better choice.OpenAI's GPT-6.1 Terra announcement, October 26
OpenAI said Terra meets the Critical threshold for cybersecurity capability under its Preparedness Framework and keeps the safeguards the company built for GPT-6 Astra.
What the launch says about OpenAI's pricing
OpenAI has now lowered the price of its Terra line twice since July. The company released the GPT-5.6 family on July 9, describing GPT-5.6 Terra as the model "for a balance of intelligence and cost," and made GPT-5.6 Terra 20% cheaper on July 30, according to its API changelog.
GPT-5.6 Terra remains available. OpenAI's deprecations page lists no shutdown date for it and still recommends it as the replacement for older models being retired, including gpt-5-mini, which shuts down on December 11.
One point OpenAI left open: it said a ChatGPT plan's allowance covers more tasks on Terra than on GPT-5.6 Terra, and substantially more than on GPT-6.1 Sol or GPT-6 Astra, without saying how many more.
