OpenAI's GPT-6.1 Terra cuts API prices 50% and targets computer use
OpenAI released GPT-6.1 Terra on Oct. 26, a cheaper mid-tier model aimed at operating computers and office work. API prices are $1 per million input tokens and $5 per million output, half of GPT-5.6 Terra's. Its benchmark claims come from OpenAI and haven't been independently checked.
AI-assisted · US Brief desk · Sources listed below
What happened
OpenAI launched GPT-6.1 Terra, a cheaper model that OpenAI says is especially strong at computer use, meaning operating software on a screen the way a person would. It is available now to Plus, Pro, Enterprise and Edu users in ChatGPT Work and Codex, and to developers as gpt-6.1-terra.
Why it matters
Cheap, reliable "computer use" is what lets companies hand routine clicking-and-typing jobs, like entering invoices into accounting software, to AI agents at scale. Halving the price makes those jobs cheaper to automate and keeps pressure on Anthropic and Google.
More context
The price: Standard API pricing is $1 per million input tokens, $0.05 per million cached input tokens and $5 per million output tokens. OpenAI says that is 50% below GPT-5.6 Terra and its larger GPT-6.1 Sol model.
OpenAI's claims: On its internal test of prompts designed to trigger mistakes, OpenAI says the error rate at low reasoning effort fell from 53.6% with GPT-5.6 Terra to 16.7%. It says Terra solves 50.3% of Terminal-Bench 4.0 coding tasks at maximum effort, versus 21.5% for the older model, and costs about a ninth as much per task as Anthropic's Claude Opus 5.5 at maximum effort. OpenAI still recommends its bigger models for the hardest coding work.
Safety note: OpenAI says Terra meets the "Critical" threshold for cybersecurity capability under its Preparedness Framework.
What's still unknown: How Terra performs in independent testing, and whether its lower error rate holds outside OpenAI's own prompt sets.
4 listed sources
References listed by US Brief; a source count is not a verification score.
Editorial sourcing notes
Sourcing: OpenAI's launch post and API model page, plus The Verge's Oct. 26 report. All benchmark and error-rate figures are OpenAI's own measurements and have not been independently verified. Hero is a 2019 file photo of the San Francisco building that housed OpenAI's offices at the time; no legitimate photo of the product exists.
Like US Brief? You can support it with a tip.