#

Key takeaways

• OpenAI introduced GPT-6.1 Terra on October 26.

• The model targets visual reasoning, computer use, and natural writing.

• OpenAI says its standard API prices are 50 percent below GPT-5.6 Terra and GPT-6.1 Sol.

• Company-reported benchmarks show gains, but they do not guarantee the same results in every workflow.

• The announcement makes no Morocco-specific availability or performance claim.

What OpenAI says GPT-6.1 Terra is for

OpenAI introduced GPT-6.1 Terra as a fast model for visual reasoning, computer use, and natural writing. The company says it is intended for common professional tasks. Those examples include understanding technical drawings, drafting documents, filling forms, and editing video.

The source is careful about how it frames those examples. It presents them as intended uses, not as independently verified deployment outcomes. That distinction matters because a model's stated purpose is not the same as proven performance in every setting.

OpenAI identifies the API model name as `gpt-6.1-terra`. It also says the model is available in ChatGPT Work and Codex for Plus, Pro, Enterprise, and Edu users. The announcement does not add any Morocco-specific availability or performance claim.

Pricing and access

OpenAI says its standard API prices for GPT-6.1 Terra are 50 percent below those of GPT-5.6 Terra and GPT-6.1 Sol. The published rates are one dollar per million input tokens, five cents per million cached input tokens, and five dollars per million output tokens.

These prices are part of the announcement, so they are the only pricing figures that can be stated here. They show a clear emphasis on lower-cost access. They do not, by themselves, describe total cost for any specific workload.

The source also links access to product tiers. It says the model is available in ChatGPT Work and Codex for Plus, Pro, Enterprise, and Edu users. That is the full access detail provided in the input.

Reported benchmark results

OpenAI reports several benchmark results for GPT-6.1 Terra. On the company's Agents' Last Exam benchmark, it says the model gained 15 percentage points over GPT-5.6 Terra at medium reasoning effort. OpenAI also says that result came at about one third of the cost.

On internal factual-error prompts, OpenAI reports a drop in error rate from 53.6 percent to 16.7 percent at low reasoning effort. On Terminal-Bench 4.0, it reports a 50.3 percent solve rate at maximum effort. The company compares that with GPT-5.6 Terra's published 21.5 percent.

These are company-reported evaluations. Their tasks and settings matter, so they do not guarantee similar results in every real workflow. That caution is important when reading benchmark claims. A strong score in one test does not automatically mean the same outcome in another environment.

How OpenAI positions the model

The announcement presents GPT-6.1 Terra as a fast option for practical work. It appears aimed at tasks that mix visual understanding, document handling, and writing. That makes the model sound broad in scope, but still focused on everyday professional use.

OpenAI also says Sol and Astra remain preferable for its most complex coding tasks. That statement helps define the model's place in the company's lineup. It suggests Terra is not the top choice for every advanced coding scenario.

The source does not provide independent verification of these claims. It also does not describe deployment outcomes from outside OpenAI. So the safest reading is that the announcement describes intended capabilities and company-reported test results.

Morocco relevance

The source reports no Morocco-specific availability or performance claim. For readers in Morocco, the global lesson is simple: check the model's stated access tier, pricing, and benchmark context before planning use.

Bottom line

GPT-6.1 Terra is positioned as a faster, lower-cost model for visual reasoning, computer use, and writing. OpenAI pairs that positioning with benchmark gains and lower API pricing. At the same time, the company's own wording keeps the claims tied to specific tests and intended use cases.