🔍 Read the full analysis: Could GPT-6.1 Terra Handle Your Work For Less? on ThorstenMeyerAI.com
Get hardware and tech essentials 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, positioning it as a lower-cost model for routine work. Artificial Analysis measured a $0.40 average task cost and a 48.1 Intelligence Index score at maximum effort; the source says low and medium settings offer the strongest cost-efficiency, while more demanding tasks may suit larger models. OpenAI’s speed and performance claims are company-reported, and independent speed measurements were not available.
OpenAI released GPT-6.1 Terra on October 26, describing it as a lower-cost model for routine communication and computer-based work. Artificial Analysis measured a $0.40 average cost per task and a 48.1 Intelligence Index score at maximum effort, placing Terra below several larger models on the benchmark while costing less to run.
Terra is positioned between OpenAI’s smaller Luna model and its larger Sol and Astra models. Its API list price is $1 per million input tokens and $5 per million output tokens, compared with $2 and $12 for its predecessor, GPT-5.6 Terra. Cached input is listed at $0.05 per million tokens. OpenAI says the model is available through the API as gpt-6.1-terra and to Plus, Pro, Enterprise and Edu users in ChatGPT Work and Codex.
Artificial Analysis evaluated Terra across ten tasks and reported its Intelligence Index score of 48.1 at maximum effort, with an average cost of $0.40 per task. In the same comparison, GPT-6.1 Sol scored 51.8 at $0.72 per task, Claude Opus 5.5 scored 57.6 at $5.98, and GPT-6 Luna scored 38.1 at $0.07. These costs represent average spending across the index’s evaluations, not a guaranteed price for every user’s task.
The source reports six effort settings, with measured scores from 33.5 to 48.1 and task costs from about $0.05 to $0.40. Low effort scored 36.2 for $0.055 per task; maximum effort cost about 7.4 times as much as low for a 12-point higher score. The source characterizes low and medium effort as lying on the cost-efficiency frontier in the tested models. It also reports that non-reasoning mode was not the cheapest setting: it cost $0.107 per task and scored 33.5.
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
When Lower Task Costs Matter
Terra’s appeal is not that it leads the benchmark, but that organizations may be able to complete recurring tasks at a lower average cost than with larger models. That could matter for high-volume workflows such as processing invoices, sorting routine requests or handling basic computer tasks. The reported figures suggest buyers should compare the cost of completing a task, including the tokens used to reach an answer, rather than relying on token rates alone.
That difference can shape how companies assign work. A lower-priced model may be suitable for predictable, lower-stakes tasks, while a more demanding request could be routed to a model with a higher benchmark score. The available figures do not establish that Terra will be cheaper for every workflow: actual cost depends on the task, prompt, output length and chosen effort setting. Organizations would need to test performance and error rates on their own work before estimating savings.
For individual users, OpenAI says the same ChatGPT Work allowance can cover substantially more tasks with Terra than with Sol or Astra. That is a company claim, not a universal measure of the allowance’s value. The trade-off is practical: spending less per task may stretch a fixed budget, but the independent scores show that Terra does not match the highest-ranked models in the cited comparison.
Products Worth Considering
Apple MacBook Pro 14-inch with M5 Max Chip (18-Core CPU, 20-Core GPU), 48GB RAM, 2TB SSD, Space Black
Apple MacBook Pro 16-inch Laptop with M5 Pro, 18-Core CPU, 20-Core GPU, 48GB Memory, 2TB SSD, Space Black
Apple MacBook Pro 14.2-inch with M5 Pro, 15-core CPU, 16-core GPU, 48GB Memory, 1TB SSD, Silver
As an affiliate, we earn on qualifying purchases.
Terra’s Place in OpenAI’s Lineup
OpenAI presents Terra as a model for the routine computer-based tasks that fill a workday, rather than as its strongest option for every use. Its launch example is reading incoming invoices and entering the information into accounting software. The company says Terra completes tasks more than three times faster than Sol and Astra, but Artificial Analysis had not published speed measurements for Terra at the time covered by the source.
OpenAI also supplied task-specific results from its own tests. On Agents’ Last Exam, it reported Terra at medium effort scoring 57.6% for $0.29 per task, compared with 42.6% and $0.84 for GPT-5.6 Terra. At maximum effort, OpenAI’s chart showed Terra at 60.0% and $1.01 per task, while Claude Opus 5.5 scored 63.2% and cost $8.82. These are OpenAI-reported measurements, and should not be confused with Artificial Analysis’s separate index results.
OpenAI also reported improvements on its internal error-inducing prompts and a 50.3% score for Terra at maximum effort on Terminal-Bench 4.0, compared with 21.5% for GPT-5.6 Terra. The company said Sol and Astra remain better choices for the most complex coding tasks. Artificial Analysis independently rated Terra at 48.1 on its Intelligence Index; its separate Coding Agent Index score was 48, below Sol’s 54 and above Luna’s 31.
Limits of the Benchmark Evidence
The figures describe benchmark performance and average costs, not results from a broad set of real-world deployments. It is not yet clear how Terra will perform across different company systems, document formats, coding projects or user prompts. Independent speed measurements were unavailable in the source, so OpenAI’s claim that Terra runs more than three times faster than Sol and Astra cannot be checked against that benchmark material.
The source also does not provide complete results for all six effort settings in the supplied text, beyond selected scores and costs. Nor does it establish the total cost or error rate for a particular company’s workflow. OpenAI’s internal factual-error and cybersecurity statements are company-reported; the source gives no independent verification of those claims. The reference to a Critical cybersecurity threshold is OpenAI’s classification under its own Preparedness Framework.
One comparison also cautions against treating Terra as the cheapest capable option: the source says Xiaomi’s open-weight MiMo-V2.6-Pro scored 46.3 for $0.13 per task, outperforming Terra’s high and extra-high settings on both score and cost in the cited data. These results are tied to the models, settings, prices and tests available on October 26, 2026, and may not reflect later changes.
Testing Terra in Real Workflows
The next useful evidence will come from independent measurements of Terra’s speed and from evaluations of how it performs on the routine office and computer tasks OpenAI is targeting. Companies considering deployment can compare its output quality, error handling and full cost against their current model on representative tasks, including the review needed when an automated system makes a mistake.
OpenAI’s launch makes Terra available through its API, ChatGPT Work and Codex for the stated eligible plans. Further model or pricing changes are not specified in the source. For now, the reported benchmarks support a limited conclusion: Terra offers a lower-cost option in several comparisons, while its measured scores and suitability vary by task and effort setting.
Key Questions
What is GPT-6.1 Terra?
GPT-6.1 Terra is an OpenAI model released on October 26, 2026, for communication and computer-based tasks. OpenAI positions it between Luna and the larger Sol and Astra models.
How much does GPT-6.1 Terra cost?
Its listed API price is $1 per million input tokens and $5 per million output tokens, with cached input at $0.05 per million tokens. Artificial Analysis measured an average cost of $0.40 per task at maximum effort in its index, but an individual task’s cost may differ.
Does Terra outperform OpenAI’s larger models?
Not across the cited independent index. At maximum effort, Terra scored 48.1, compared with 51.8 for Sol and 52.7 for Astra in the supplied comparison. OpenAI reports stronger results for Terra on some specific tests, but those are the company’s own measurements.
Why compare cost per task instead of token price?
Token price does not show how many tokens a model uses to complete a task. Artificial Analysis’s cost-per-task estimate includes spending across its evaluation tasks, so it can produce a different ranking from token rates alone. It is still an average benchmark measure, not a quote for every workflow.
What has not been independently verified?
The source says Artificial Analysis had not published speed measurements for Terra. OpenAI’s speed, internal error-rate and cybersecurity statements are company-reported, and the available benchmark data does not establish how Terra will perform on each organization’s real work.
Source: ThorstenMeyerAI.com
