🔍 Read the full analysis: Why GPT-6.1 Terra Might Be The Model To Use When Costs Add Up on ThorstenMeyerAI.com
Get business pricing on tech for your team
- Business-only prices and quantity discounts
- Tax-exempt purchasing
- Multiple users, one account, clear invoices
TL;DR
OpenAI released GPT-6.1 Terra on October 26, 2026, pricing API use at $1 per million input tokens and $5 per million output tokens. Artificial Analysis found Terra’s low and medium effort settings on its cost-efficiency frontier, but higher scores remain available from more capable models, and some cheaper alternatives also scored well.
OpenAI released GPT-6.1 Terra on October 26, pricing API use at $1 per million input tokens and $5 per million output tokens—half the listed rates for its predecessor, GPT-5.6 Terra. Independent testing by Artificial Analysis found that Terra’s low and medium reasoning settings sit on its cost-efficiency frontier, but the model does not lead the firm’s intelligence rankings.
Artificial Analysis gave Terra a score of 48.1 on its Intelligence Index at maximum effort, below GPT-6.1 Sol’s 51.8 and Claude Opus 5.5’s 57.6. Its average cost across the index’s evaluation tasks was $0.40 per task, compared with $0.72 for Sol and $5.98 for Opus 5.5. These figures describe the benchmark’s task mix, not a guaranteed price for a particular customer’s workload.
Terra has six effort settings, which affect both results and cost. Artificial Analysis measured scores ranging from 33.5 at non-reasoning to 48.1 at maximum effort, with average task costs from roughly $0.05 to $0.40. The source reports that maximum effort costs 7.4 times as much as low effort for 12 additional index points. Low and medium settings were on the firm’s cost-efficiency frontier; settings above them were not.
OpenAI says Terra is intended for frequent communication and computer-based work, such as processing invoices. It is available through ChatGPT Work and Codex for Plus, Pro, Enterprise and Edu users, as well as through the API under the model name gpt-6.1-terra. OpenAI says the same allowance can cover substantially more tasks than with Sol or Astra; that is the company’s claim, not an independent measurement of every user’s spending.
Infographic 01 · The effort curve
One model, six settings, a 7-fold cost range.
GPT-6.1 Terra at each effort level, as measured by Artificial Analysis. Low and medium sit on its cost-efficiency frontier: no tested model is both cheaper and better.
Score · Intelligence Index
independent higher is better
Cost per finished task
note the last step shorter is cheaper
All six settings
| Effort | Index | Cost / task | Output tokens / task | Hallucination rate |
|---|---|---|---|---|
| Non-reasoning | Index 33.5 | Cost $0.107 | Tokens 4,737 | Hallucination 72% |
| Low | Index 36.2 | Cost $0.055 | Tokens 2,681 | Hallucination 46% |
| Medium | Index 41.3 | Cost $0.095 | Tokens 5,835 | Hallucination 44% |
| High | Index 43.9 | Cost $0.134 | Tokens 9,027 | Hallucination 43% |
| Xhigh | Index 45.4 | Cost $0.194 | Tokens 14,551 | Hallucination 42% |
| Max | Index 48.1 | Cost $0.403 | Tokens 35,096 | Hallucination 38% |
Hallucination rate in AA-Omniscience: how often the model answers anyway when it does not know. Lower is better.
The first steps up are cheap. The last one doubles the bill.Assessment: Thorsten Meyer
Sources: Artificial Analysis model pages (retrieved 26 Oct 2026); Artificial Analysis launch article (26 Oct 2026). Cost per task is the average across the index’s evaluation tasks.
Powered by Thorsten Meyer AI
Why Per-Task Costs Matter
Token prices are only one part of the cost of using a model. A model may charge less for each token but use more tokens to complete a request; the resulting bill depends on the task, its prompts, the model’s output and any reasoning tokens. Artificial Analysis’s comparison therefore matters because it estimates average spending per completed benchmark task, rather than ranking models by token rates alone.
The results suggest Terra could suit organizations that handle many routine requests and can route difficult work elsewhere. Artificial Analysis found low and medium settings more cost-efficient on its index, while OpenAI itself says Sol and Astra remain better choices for the most complex coding tasks. That points to a practical trade-off: lower cost for suitable workloads, not a universal replacement for more capable models.
Terra is also not the sole lower-cost option in the published comparison. Artificial Analysis lists open-weight MiMo-V2.6-Pro at 46.3 points and $0.13 per task, better on both measures than Terra at its high and extra-high settings. Buyers should compare models on their own workloads, including accuracy, latency, privacy and operational costs, rather than treating a single index as a purchasing decision.
Products Worth Considering
Samsung 27-Inch Odyssey QD-OLED G8 4K 240Hz Gaming Monitor
LG 34GX90SA-W 34-inch Ultragear WQHD OLED Curved Gaming Monitor
Samsung 27 Odyssey OLED G5 (G50SF) QHD QD-OLED Gaming Monitor
As an affiliate, we earn on qualifying purchases.
Terra’s Place in OpenAI’s Lineup
OpenAI positions Terra between the smaller GPT-6 Luna and the larger Sol and Astra models. The company describes it as a fast, efficient option for work that happens repeatedly across a business. Its listed API price is $1 per million input tokens and $5 per million output tokens, with cached input priced at $0.05 per million tokens. Artificial Analysis lists Luna’s token rates at $0.10 input and $0.50 output per million tokens, illustrating why a lower per-token price does not by itself establish which model costs less to finish a task.
OpenAI also published its own task evaluations. On Agents’ Last Exam, it reports Terra at medium effort scoring 57.6%, compared with 42.6% for GPT-5.6 Terra at medium, and costing $0.29 rather than $0.84 per task. On Terminal-Bench 4.0, OpenAI reports a 50.3% result at maximum effort, compared with 21.5% for the predecessor’s published result. These are company-reported results and should not be conflated with Artificial Analysis’s separate index.
Artificial Analysis says Terra’s maximum-effort score is six points above GPT-5.6 Terra’s while costing less than a third as much per benchmark task: $0.40 versus $1.40. It also reports that Terra guessed on 38% of questions it could not answer in its hallucination test, compared with 88% for GPT-5.6 Terra. Those are specific test results, not proof that the model will avoid errors in other settings.
Limits of the Published Benchmarks
The published figures do not establish how Terra will perform across every customer’s software, data or task mix. Artificial Analysis’s cost-per-task figure is an average across its evaluation set; a company’s actual bill may differ depending on prompt length, output size, reasoning setting, caching and repeated attempts. The supplied source also says the firm had not published speed measurements for Terra at the time of its review, so OpenAI’s claim that Terra completes tasks more than three times faster than Sol and Astra had no corresponding independent speed figure in that material.
OpenAI reports lower error rates on its internal error-inducing prompts and classifies Terra as meeting its Preparedness Framework’s Critical threshold in cybersecurity. Both statements come from OpenAI, and the source does not provide enough detail to independently assess the test design or classification. It also remains unclear how much the benchmark advantages will translate into savings after integration, human review and error correction are included.
How Teams Can Test Terra
The next useful evidence will come from independent speed measurements and customer results on real workloads. Teams considering Terra can compare low and medium settings against their current models using representative tasks, tracking cost per accepted result, accuracy, latency and the amount of human correction required. Testing only token prices or a public benchmark score would leave important parts of the operating cost unmeasured.
For now, the launch establishes Terra as a lower-priced option in OpenAI’s lineup, with independent results favoring particular effort settings rather than the model in every configuration. Whether it saves money depends on the work being assigned to it—and whether its output is good enough to use without costly revisions.
Key Questions
How much does GPT-6.1 Terra cost?
OpenAI lists API pricing at $1 per million input tokens and $5 per million output tokens. Cached input is listed at $0.05 per million tokens. Actual task costs vary with usage and settings.
Is GPT-6.1 Terra OpenAI’s highest-scoring model?
No. Artificial Analysis scored Terra at 48.1 at maximum effort, below GPT-6.1 Sol at 51.8 and Claude Opus 5.5 at 57.6 on its Intelligence Index.
Which Terra settings had the best cost-efficiency results?
Artificial Analysis placed Terra’s low and medium effort settings on its cost-efficiency frontier. The source reports that higher settings cost more without offering the same efficiency trade-off on that index.
Does a cheaper token price mean a cheaper completed task?
No. A task’s total cost also depends on how many input, output and reasoning tokens the model uses. Artificial Analysis’s results show why comparing cost per completed task can give a different picture from comparing token prices alone.
Where can users access GPT-6.1 Terra?
According to the source, Terra is available in ChatGPT Work and Codex for Plus, Pro, Enterprise and Edu users, and through the API as gpt-6.1-terra.
Source: ThorstenMeyerAI.com
