🔍 Read the full analysis: GPT-6.1 Terra Could Make Routine AI Work More Affordable 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 it at $1 per million input tokens and $5 per million output tokens—half the listed rates for GPT-5.6 Terra. Artificial Analysis found Terra’s low- and medium-effort settings on its cost-efficiency frontier, but higher effort costs more and some competing models deliver stronger scores or lower task costs.
OpenAI released GPT-6.1 Terra on October 26, pricing it at $1 per million input tokens and $5 per million output tokens, half the listed rates for GPT-5.6 Terra. The model is aimed at routine workplace and computer-based tasks; independent testing by Artificial Analysis found its low- and medium-effort settings offered the strongest cost-efficiency among Terra’s six settings, while the highest setting cost more for a higher score.
Terra is available in ChatGPT Work, Codex and the API, where the model identifier is gpt-6.1-terra. OpenAI describes it as a fast, efficient model for communication and computer-based work, citing invoice reading and data entry as examples. The company says Plus, Pro, Enterprise and Edu users can complete substantially more tasks under the same allowance than with Sol or Astra. That allowance claim comes from OpenAI; the source material does not provide a measured allowance comparison.
Artificial Analysis reported a score of 48.1 on its Intelligence Index for Terra at maximum effort, at an average cost of $0.40 per task across the index’s evaluation tasks. GPT-6.1 Sol scored 51.8 at $0.72 per task, while Claude Opus 5.5 scored 57.6 at $5.98. These results show Terra is cheaper per task than those models in this benchmark, but not the highest-scoring model. The index combines ten evaluations, and its task-cost figure is an average across them.
Artificial Analysis measured Terra across six effort settings, with scores ranging from 33.5 at non-reasoning to 48.1 at maximum effort and average task costs from $0.055 to $0.40. Its results put low and medium effort on the cost-efficiency frontier: no other tested model in the comparison was both cheaper and better at those points. The source reports that maximum effort costs 7.4 times as much as low effort for 12 additional index points. The non-reasoning setting was not the cheapest, costing almost 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
The pricing change matters most to organizations running large volumes of repeatable AI work. A lower rate per token can reduce bills, but it does not by itself show how much a completed job costs. Artificial Analysis’s results reflect the number of tokens each model uses as well as its listed rates: GPT-6 Luna, for example, is listed at $0.10 per million input and output tokens, yet Terra at low effort had a lower average cost per task in this evaluation. The practical comparison is the cost of a finished task, alongside its quality and reliability.
For teams, Terra’s effort options create a budgeting choice. Low or medium effort may suit routine work where the benchmark’s trade-off is acceptable; more demanding tasks may justify greater effort or a more capable model. Artificial Analysis’s results do not prove Terra will be the cheapest option for every company’s workload. Costs can change with task type, prompt length, output length and usage patterns, so organizations will need to test their own jobs before setting budgets or routing rules.
There are also limits to what the scores establish. Artificial Analysis’s index is a benchmark, not a measurement of a particular company’s productivity or return on investment. OpenAI reports improvements in computer use, coding and factual-error tests, but those are company-reported measurements and should not be treated as independent validation. The release therefore gives buyers a new price and model option, not a guarantee of savings in every deployment.
Products Worth Considering
Jackery Explorer 300 Portable Power Station with 40W Solar Panel
Bose S1 Pro+ Portable PA Speaker
Samsung 27-inch Odyssey OLED G6 QD-OLED QHD Gaming Monitor
As an affiliate, we earn on qualifying purchases.
How Terra Fits OpenAI’s Lineup
OpenAI positions Terra between GPT-6 Luna, its smaller model, and the larger Sol and Astra models. The launch material says Terra’s API prices are 50% lower than GPT-5.6 Terra’s and GPT-6.1 Sol’s. Its listed price is also below the figures given for Sol and Astra in the Artificial Analysis comparison. The source does not establish that Terra replaces those models; OpenAI says Sol and Astra remain better choices for the most complex coding tasks.
Artificial Analysis’s launch-day comparison places Terra below several higher-scoring models at maximum effort, including Claude Opus 5.5, Kimi K3.1 Pro and Gemini 4 Argon. It also shows lower-cost competitors: Xiaomi’s open-weight MiMo-V2.6-Pro scored 46.3 at an average task cost of $0.13, compared with Terra’s maximum-effort score of 48.1 at $0.40. These figures are specific to the index and its listed settings, rather than a universal ranking across all uses.
OpenAI also reports that Terra completed tasks more than three times faster than Astra and Sol, and says it meets the Critical threshold in cybersecurity under the company’s Preparedness Framework. Those statements are OpenAI’s own claims. The source says Artificial Analysis had not published speed measurements for Terra when its figures were checked on October 26, 2026.
Limits of the Launch-Day Evidence
It is not yet clear how Terra’s benchmark results will translate to real-world workloads, where task difficulty, context size and error tolerance vary. Artificial Analysis reports average costs across its evaluation tasks; those values do not predict every user’s bill. The source also does not give a complete breakdown of the six effort settings’ performance on each individual evaluation.
OpenAI’s speed, factual-error, computer-use and coding comparisons are based on company-reported tests, and the supplied material does not include independent confirmation of those claims. The launch-day source also reports that independent speed measurements were not available. Terra’s cybersecurity classification is likewise OpenAI’s assessment under its own framework, not an external finding described in the material.
The source says Terra guessed on 38% of questions it could not answer in Artificial Analysis’s hallucination test, down from 88% for GPT-5.6 Terra. That comparison is reported as a rate of guessing on unanswerable questions; it does not mean Terra is error-free or establish how often it will make mistakes in workplace use. The supplied six-setting data is also incomplete, so detailed figures for every setting cannot be stated here.
Testing Terra on Workflows
The immediate next step for prospective users is to test Terra against their own routine tasks and compare cost per completed job, output quality and error rates across effort settings. For high-volume deployments, teams can start by evaluating low or medium effort, then determine whether the additional cost of higher effort produces enough improvement for harder tasks. OpenAI’s availability in ChatGPT Work, Codex and the API gives customers several routes to try the model.
Further independent measurements could clarify Terra’s speed and performance across specific task types, while real usage data will show whether the benchmark’s cost pattern holds for business workloads. Until then, the confirmed development is the model’s release and listed pricing; the scale of any customer savings remains dependent on how Terra performs on each customer’s work.
Key Questions
How much does GPT-6.1 Terra cost?
OpenAI lists API prices of $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. Actual task costs depend on token use and the effort setting.
Where is GPT-6.1 Terra available?
The model is available in ChatGPT Work and Codex for Plus, Pro, Enterprise and Edu users, and through the API as gpt-6.1-terra, according to the launch material.
Which Terra effort setting is cheapest?
Artificial Analysis measured low effort at an average of $0.055 per task, the lowest cost among the Terra settings reported in the source. Non-reasoning was not the cheapest and cost almost twice as much per task as low effort.
Is Terra the strongest model in the comparison?
No. At maximum effort, Artificial Analysis scored Terra at 48.1, below Claude Opus 5.5, Kimi K3.1 Pro, Gemini 4 Argon, GPT-6 Astra and GPT-6.1 Sol in the comparison. Terra’s case is its lower task cost in some settings, not the highest benchmark score.
Will Terra reduce a company’s AI costs?
That is not established for every company. Independent benchmark results suggest low- and medium-effort Terra can be cost-efficient on the tested tasks, but actual savings depend on the workload, token use, accuracy needs and the alternative model being compared.
Source: ThorstenMeyerAI.com
