1. What was announced
On October 26, OpenAI announced GPT-6.1 Terra, a model that combines detailed visual reasoning and precise computer use with clearer, more natural writing. OpenAI says it handles everyday work such as drafting proposals, interpreting technical drawings, filling out forms and editing videos. Standard API pricing is $1 per million input tokens, $0.05 per million cached input tokens and $5 per million output tokens, which OpenAI says is 50% lower than GPT-5.6 Terra and GPT-6.1 Sol (OpenAI announcement).
It is available from the same day to Plus, Pro, Enterprise and Edu users in ChatGPT Work and Codex, and through the API as gpt-6.1-terra. OpenAI also says the same usage allowance covers more tasks than with GPT-5.6 Terra. By OpenAI's own evaluation, on the computer-use test Agents' Last Exam it scores 15 percentage points higher than GPT-5.6 Terra at medium reasoning effort at about one-third of the cost, and finishes tasks over three times faster than GPT-6 Astra and GPT-6.1 Sol. On error-inducing prompts, the error rate at low reasoning effort fell from 53.6% to 16.7%. All of these figures are OpenAI's own measurements.
BleepingComputer reported the same details the same day (BleepingComputer), and in Japanese, Mado no Mori reported on October 27 that API pricing is half that of the predecessor (Mado no Mori via Yahoo News Japan). OpenAI states that GPT-6.1 Sol and GPT-6 Astra remain the better choice for the most complex coding tasks, so Terra is positioned as the balance of capability and cost.
2. Why it matters to small businesses
As an example use, OpenAI cites high-volume routine work such as reading incoming invoices and entering their details into accounting software, and says the lower price makes automating it more practical. For a company where staff re-key dozens of invoices or purchase orders each month, a lower cost per document changes the calculation of whether to hand the work to AI. OpenAI also states that the model meets the Critical threshold in cybersecurity under its Preparedness Framework, so it is worth remembering that the same capability can be used by attackers too.
Our view What has changed with this announcement is less how smart the AI is than what it costs per item to hand routine work to it. Invoice processing or order entry that did not pay off six months ago may pay off now. If a competitor lowers its back-office cost first, the gap will show up in pricing and lead times, and a company that keeps processing by hand will be at a disadvantage when quotes are compared. List your monthly routine tasks by volume and time spent, and set them beside the estimated cost of automating them. Treat OpenAI's own figures as a guide only, and decide after testing on your own documents.
3. What small businesses should check now
- Write down the routine work you repeat each month (keying invoices, entering orders, first replies to inquiries) by volume and time spent
- Ask your providers whether GPT-6.1 Terra is available under your current AI contracts (ChatGPT plan, API, or features built into accounting and order systems)
- Test reading and entry on about ten of your own documents and record how errors appear; OpenAI's error rates come from its internal test and may not match your documents
- Decide in an internal rule what information may be given to AI, such as customers' personal data and price lists
- Name the person who does the final check when the AI gets something wrong
Our outlook
Over the next six months to two years, we expect AI usage prices to keep falling, and back-office work such as invoice processing and order entry to shift from work people do to work AI does and people check. In the same period, AI will increasingly request quotes, compare and place orders on behalf of consumers and buyers at other companies. Companies that publish products, prices, stock and lead times in a machine-readable form and can answer AI inquiries immediately will stay on the shortlist; others are likely to be dropped at the comparison stage. Automating back-office work and making yourself visible to AI are two sides of the same preparation. Now is the time to lay that groundwork, before competitors and their AI agents do.
Questions this article answers
Is GPT-6.1 Terra really cheaper?
According to OpenAI, standard API pricing is $1 per million input tokens and $5 per million output tokens, 50% lower than GPT-5.6 Terra and GPT-6.1 Sol. In ChatGPT Work and Codex, the same usage allowance is also said to cover more tasks. Actual cost depends on volume, so run an estimate on your own documents.
What should we prepare to automate invoice processing with GPT-6.1 Terra?
First write down how many invoices you handle each month and how long they take, then test reading and entry on about ten real documents and record the errors. Next decide what information may be given to AI, and name a person to do the final check. OpenAI cites reading invoices and entering them into accounting software as an example use, but its error rates come from its own internal test.
Sources
- OpenAI: Introducing GPT-6.1 Terraopenai.com
- BleepingComputer: OpenAI launches GPT-6.1 Terra with 3x faster AI agents, 50% lower pricingbleepingcomputer.com
- 窓の杜（Yahoo ニュース）: OpenAI、「GPT-6.1 Terra」を発表 ～性能とコストのバランス型、API料金は先代の半額news.yahoo.co.jp
Related Q&A
Related services
If you want to try this in your own business
To see how the topics in this article would apply to your own operations, and what it would cost, send us a message through the contact form.