🔍 Read the full analysis: Could GPT-6.1 Terra Save Money On AI Without Changing Your Workflow? on ThorstenMeyerAI.com
Get the latest gadgets delivered free with Prime
- Fast, free delivery on millions of items
- Prime Video, Amazon Music and more included
- Member-only deals all year
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 effort raises costs and scores below leading models. Actual savings depend on the task, usage and chosen model.
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 per-token prices of its predecessor, GPT-5.6 Terra. The launch positions Terra as a lower-cost option for routine work, but independent tests show that whether it saves money depends on its reasoning setting and the cost of completing a task—not token prices alone.
Terra is available through the API as gpt-6.1-terra and to Plus, Pro, Enterprise and Edu users in ChatGPT Work and Codex. OpenAI describes it as a fast, efficient model for communication and computer-based work, citing tasks such as reading incoming invoices and entering information into accounting software. The company says its paid ChatGPT plans provide substantially more tasks with Terra than with Sol or Astra, though it did not quantify that allowance claim in the supplied material.
Artificial Analysis, which published an evaluation on launch day, gave Terra a 48.1 Intelligence Index score at maximum effort and estimated an average cost of $0.40 per task across its benchmark tasks. By comparison, its listed figures were $0.72 for GPT-6.1 Sol and $5.98 for Claude Opus 5.5. Opus scored 57.6, however, compared with Terra’s 48.1. Those figures reflect the index’s tests and cost accounting, not a guarantee of what any particular customer will pay.
The six-setting curve makes the trade-off clearer: Artificial Analysis measured scores from 33.5 at non-reasoning to 48.1 at maximum effort, while average task cost ranged from $0.055 at low effort to $0.40 at maximum. Low and medium effort were on the firm’s cost-efficiency frontier—meaning no other model it tested was both cheaper and better at those points. The supplied source says maximum effort costs 7.4 times as much as low effort for 12 additional index points. It also reports that non-reasoning cost nearly twice as much per task as low effort while scoring 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
Task Costs Shape Terra’s Value
Terra’s launch matters to organizations buying AI at high volume because per-token prices can mislead when models generate different amounts of text or take different paths to finish a task. In Artificial Analysis’s figures, GPT-6 Luna costs one-tenth as much as Terra per token, but the source reports that Terra at low effort costs less per completed benchmark task because it uses fewer tokens. Buyers comparing models on price alone could miss that difference.
For companies, the practical question is whether Terra can handle a defined workload reliably at an effort setting that keeps total costs down. OpenAI’s invoice-processing example points to repetitive work where task volume can make small unit-cost differences add up. But the available evidence does not establish savings for every deployment: benchmark averages may not match a company’s prompts, software connections, review requirements or error costs. A cheaper run is not necessarily cheaper overall if people must correct more mistakes.
Terra also gives users a choice between spending more for a stronger result and limiting costs on routine work. Artificial Analysis measured higher scores at higher effort, but the cost rose along with them. OpenAI says Sol and Astra remain better choices for its most complex coding tasks. That makes Terra a potential fit for some workloads, not a like-for-like replacement for every larger model.
Products Worth Considering
LG UltraGear 32GX850A-B 32-Inch 4K UHD Glossy OLED Gaming Monitor
Samsung 32-Inch Odyssey OLED G8 (G80SH) 4K Gaming Monitor
LG 34GX90SA-W 34-inch Ultragear WQHD OLED Curved Gaming Monitor
As an affiliate, we earn on qualifying purchases.
How Terra Fits OpenAI’s Lineup
OpenAI places Terra between its smaller Luna model and larger Sol and Astra models. The launch page presents Terra as a model for everyday communication and computer-based tasks, while the supplied source describes Sol and Astra as preferable for more complex coding. This makes Terra’s appeal partly about matching model capability to task difficulty, rather than selecting one model for all uses.
The company reported improvements on several of its own evaluations. On the computer-use benchmark Agents’ Last Exam, OpenAI said Terra at medium effort scored 57.6%, compared with 42.6% for GPT-5.6 Terra, and cost $0.29 rather than $0.84 per task; it also said that result matched GPT-6 Astra at medium effort, which cost $4.82 per task. On Terminal-Bench 4.0, OpenAI reported that Terra at maximum effort solved 50.3% of tasks, versus 21.5% for GPT-5.6 Terra in its published result. These are company-reported benchmark results, separate from Artificial Analysis’s independent evaluation.
Artificial Analysis’s broader comparison also shows that Terra is not the top-scoring model in its listed set. Claude Opus 5.5, Kimi K3.1 Pro and Gemini 4 Argon all scored higher on the firm’s index. The source also lists Xiaomi’s open-weight MiMo-V2.6-Pro at 46.3 for $0.13 per task, below Terra’s maximum-effort score and cost. These comparisons underline that Terra competes on a particular price-performance curve, rather than leading every category.
Real-World Savings Remain Untested
The benchmark figures do not show what an individual business will pay for its own workload. Actual task costs vary with prompts, output length, cached input, reasoning setting and the number of retries or human checks required. Artificial Analysis’s $0.40 figure is its average across index tasks, not a fixed charge for every Terra request.
Several performance and safety statements in the launch material come from OpenAI rather than an independent assessment. The source says OpenAI reported Terra completes tasks more than three times faster than Astra and Sol, but Artificial Analysis had not published speed measurements for Terra when the figures were checked on October 26. OpenAI also said Terra meets the Critical threshold in its cybersecurity Preparedness Framework and retains safeguards developed for GPT-6 Astra; those are the company’s own classifications.
The supplied material does not establish how Terra will perform across different customer workflows, whether OpenAI’s reported allowance increase will translate into lower total bills, or how benchmark error rates compare in live deployments. Artificial Analysis’s evaluation offers one basis for model comparison, but it cannot resolve those questions for every use case.
Compare Models on Your Workload
Terra is available now through the API, ChatGPT Work and Codex for the listed eligible plans. Organizations considering it can compare cost per completed task across low, medium and higher effort settings using representative work, then check accuracy and the need for human review. Tests should include the full workflow cost, not just the model’s token charge.
Further independent measurements could clarify how Terra performs on speed and other real-world workloads. Until then, the clearest next step for buyers is to treat OpenAI’s launch figures as company-reported and Artificial Analysis’s task averages as benchmark-specific. Whether Terra saves money without workflow changes remains an outcome to test, rather than a confirmed result for all users.
Key Questions
How much does GPT-6.1 Terra cost?
OpenAI’s listed API price is $1 per million input tokens and $5 per million output tokens. Cached input is listed at $0.05 per million tokens in the source material. Total costs depend on usage and the task.
Does Terra save money compared with Luna?
Not necessarily on token price: the source lists Luna at one-tenth Terra’s per-token price. Artificial Analysis’s benchmark figures suggest Terra at low effort can cost less per completed task because it uses fewer tokens. That comparison may not hold for every workload.
Which Terra effort setting is cheapest?
Artificial Analysis measured low effort at $0.055 per task on average, the lowest cost among the settings described. Its non-reasoning setting cost $0.107 on average and scored lower, according to the source.
Is Terra better than OpenAI’s larger models?
Not across the measures provided. Artificial Analysis scored Terra below Sol and Astra in its listed comparison, and OpenAI says Sol and Astra remain better choices for its most complex coding tasks. Terra’s case is lower cost for selected tasks, not top performance across the board.
Are the reported savings confirmed for business users?
No. The available cost figures are benchmark averages, while OpenAI’s allowance and performance statements are company claims. Businesses would need to test Terra on their own tasks and include error correction and review costs.
Source: ThorstenMeyerAI.com
