OpenAI released GPT-6.1 Terra, a mid-range model. Here is where a team that automates repetitive work could use it.
GPT-6.1 Terra is an AI model that OpenAI released this week. OpenAI's announcement describes it as a fast, efficient model that combines detailed visual reasoning and precise computer use with clearer, more natural writing. It calls the model a step up from GPT-5.6 Terra.
Computer use is a model's ability to look at a screen, click and type to get work done. Filling out forms and editing videos are among the examples.
The announcement scored 597 points on Hacker News. On 2026-10-26, Simon Willison pointed out what the announcement put first: visual reasoning and computer use.
Image: Simon Willison
According to Simon Willison's post, GPT-6.1 Terra costs 1 dollar per million input tokens and 5 dollars per million output tokens. That is almost half the price of GPT-5.6 Terra and GPT-6.1 Sol, and about 10 times the price of GPT-6 Luna.
OpenAI's current lineup is GPT-6 Astra, GPT-6.1 Sol, GPT-6.1 Terra and GPT-6 Luna. Terra sits where it is cheaper than Sol and smarter than Luna. Simon Willison compared that spot to where Sonnet sits in Anthropic's lineup.
Image: OpenAI
Simon Willison shared a Hacker News comment by OpenAI's Ted Sanders. In summary:
- For technical writing that has to be precise, it is better than Astra and Sol
- On vision and computer use it is similar, but much faster
- On most other things it is a bit worse
For API developers it is a mid-range option: cheaper than Sol and a lot smarter than Luna. For subscribers, he says it is not a big change. If you keep hitting usage limits it may help you go a little further, and if you want clearer writing or faster computer use it is worth picking.
For a team that automates repetitive work, computer use is why this model stands out. Simple, repeated tasks such as filling the same form every day do not need the most expensive model. Speed and price matter more. Simon Willison makes a similar point.
Still, the comparison above is an OpenAI employee's account and one blogger's summary. It is not a result on our own work. So set an order before switching.
- Pick one repetitive task that the model you use now handles
- Run the same task on Terra and write the cost per task and the time it takes side by side
- Look at how it fails: does it misread the screen, or skip steps
- On GitHub trending (TypeScript), ruvnet/ruflo gained 2,683 stars in the week
- Also on GitHub trending, anthropics/knowledge-work-plugins gained 2,822 stars in the week
- OpenAI published a case on how Fanatics streamlines the fan experience
Sources
- OpenAI, Introducing GPT-6.1 Terra
- Simon Willison, GPT-6.1 Terra, 2026-10-26