🔍 Read the full analysis: A Different Kind Of Value In OpenAI’s GPT-6.1 Terra on ThorstenMeyerAI.com
Get business pricing on office and shipping supplies
- 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 measured Terra’s cost per task across six effort settings; it found low and medium settings on its cost-efficiency frontier, while higher settings cost more for gains in benchmark scores. Terra is not the highest-scoring model in the supplied comparison.
OpenAI released GPT-6.1 Terra on October 26, a model positioned between its smaller Luna and larger Sol and Astra models, with API prices of $1 per million input tokens and $5 per million output tokens. Independent testing by Artificial Analysis found that Terra’s low and medium reasoning settings offer a strong cost-to-score balance on its benchmark, though Terra does not lead the overall scores.
OpenAI describes Terra as a fast, efficient model for communication and computer-based work, including high-volume tasks such as processing invoices. It is available through the API under the name gpt-6.1-terra and to Plus, Pro, Enterprise and Edu users in ChatGPT Work and Codex. OpenAI says its subscription allowance can cover substantially more Terra tasks than comparable use of Sol or Astra; that is the company’s description, not an independently verified allowance comparison.
Artificial Analysis gives Terra a score of 48.1 on its Intelligence Index at maximum effort, at an average cost of $0.40 per task across the index’s evaluations. In the same comparison, GPT-6.1 Sol scored 51.8 at $0.72 per task, while Claude Opus 5.5 scored 57.6 at $5.98. These figures are benchmark averages, not a fixed price for every real-world task. Artificial Analysis says Terra’s low and medium effort settings fall on its cost-efficiency frontier; its higher settings do not.
Terra offers six effort settings, trading higher benchmark scores for higher average task costs. Across the full range, its index score rises from 33.5 in non-reasoning mode to 48.1 at maximum effort, while the reported average cost rises from $0.107 to $0.40 per task. Low effort scored 36.2 and cost $0.055 per task. The supplied source says maximum effort costs 7.4 times as much as low effort for 12 more points, while the non-reasoning setting costs almost twice as much per task as low effort and scores lower.
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 Task Costs Change the Comparison
Terra’s launch highlights a distinction between token pricing and the cost of finishing a task. A model with a lower price per token may use more tokens to complete a job, while another may produce a shorter answer or need less reasoning. Artificial Analysis’s task-cost measure includes spending across its evaluation set, so it offers a different basis for comparing models than the API price sheet alone.
That distinction could matter to organizations running repeated workflows, where cost per completed job affects how many tasks a budget can cover. Terra’s low and medium settings appear to offer the clearest value in the cited independent results. But the benchmark does not establish that Terra will be cheaper or more effective for every company’s workload; performance and token use can vary by task.
The figures also show a trade-off rather than a simple replacement for larger models. Terra’s maximum-effort score trails Sol and several other systems in Artificial Analysis’s comparison. OpenAI says Sol and Astra remain better choices for the most complex coding work. For users, the practical question is whether a task can be handled at a lower effort setting or requires a higher-scoring model.
Products Worth Considering
Dell UltraSharp 27 Thunderbolt Hub Monitor (U2726DER)
Dell UltraSharp 32 U3226QER 4K Thunderbolt Hub 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 places Terra between Luna, its smaller model, and Sol and Astra. Its API prices are half those listed for GPT-5.6 Terra: the new model costs $1 per million input tokens and $5 per million output tokens, compared with $2 and $12 for the predecessor. Cached input is listed at $0.05 per million tokens.
OpenAI’s own launch-page charts report gains on selected tests. At medium effort, Terra scored 57.6% on Agents’ Last Exam, compared with 42.6% for GPT-5.6 Terra, and the company lists respective task costs of $0.29 and $0.84. At maximum effort, OpenAI reports a 50.3% score on Terminal-Bench 4.0, against a published 21.5% for GPT-5.6 Terra. These are company-provided benchmark results, separate from Artificial Analysis’s independent index.
Artificial Analysis 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. OpenAI reports lower error rates on a separate internal set of prompts. The tests use different methods and should not be treated as interchangeable evidence of performance across all uses.
Limits of the Published Comparisons
The available figures do not establish how Terra performs across every production workload. Artificial Analysis’s cost per task is an average across its own evaluations, and actual costs will depend on task type, effort setting and token use. The supplied material also does not provide all six effort-setting scores and costs in a complete table, so a full setting-by-setting comparison cannot be reproduced here.
Speed is another open point: OpenAI says Terra completes tasks more than three times faster than Astra and Sol, but Artificial Analysis had not published speed measurements for Terra at the time reported. OpenAI also says Terra meets its Critical cybersecurity threshold under its Preparedness Framework and retains safeguards developed for Astra; that classification is OpenAI’s own assessment.
The results do not show whether users will see the same allowance advantage in ChatGPT Work or Codex across different plans and workloads. Nor do the reported benchmark scores settle how Terra compares with other models on a particular customer’s tasks.
What Users Can Measure Next
As companies and individual users try Terra, the most useful comparisons will track cost and success per completed task, alongside response quality and time to completion. Teams considering automated workflows can test representative jobs at low and medium effort before deciding whether higher settings or a larger model are needed.
Further independent speed measurements and evaluations on a wider range of tasks would help test OpenAI’s performance claims. For now, the launch-day evidence supports a narrower conclusion: Terra offers lower listed token prices than its predecessor and strong cost-efficiency results at selected effort settings, but it is not the top-scoring model in the cited benchmark comparison.
Key Questions
What is GPT-6.1 Terra?
GPT-6.1 Terra is a new OpenAI model positioned between Luna and the larger Sol and Astra models. OpenAI says it is designed for communication and computer-based work.
How much does GPT-6.1 Terra cost?
OpenAI lists API prices of $1 per million input tokens and $5 per million output tokens, with cached input at $0.05 per million tokens. Costs per task vary with the task and effort setting.
Is Terra the highest-scoring model in the reported comparison?
No. Artificial Analysis scored Terra at 48.1 at maximum effort. Its comparison lists higher scores for Claude Opus 5.5, Kimi K3.1 Pro, Gemini 4 Argon, GPT-6 Astra and GPT-6.1 Sol.
Which Terra effort setting offers the strongest cost-efficiency?
Artificial Analysis places low and medium effort on its cost-efficiency frontier. That finding applies to its benchmark and should not be assumed to hold for every real-world task.
Where can people use GPT-6.1 Terra?
Terra is available through the API as gpt-6.1-terra and to Plus, Pro, Enterprise and Edu users in ChatGPT Work and Codex, according to the launch material.
Source: ThorstenMeyerAI.com
