Artificial Intelligence
OpenAI releases GPT-6.1 Terra at half API price
OpenAI said GPT-6.1 Terra was a mid-range model for routine office work and software control, priced at half the rate of GPT-5.6 Terra and sold through ChatGPT Work, Codex and the API.
- OpenAI released GPT-6.1 Terra and said its API prices were half those of GPT-5.6 Terra.
- The company said Terra was available to Plus, Pro, Enterprise and Edu users in ChatGPT Work and Codex, as well as through the API.
- Artificial Analysis said Terra improved the Intelligence Index and cost less than a third per task at maximum reasoning effort.
- OpenAI said Terra met its "Critical" cybersecurity threshold and had a higher indirect prompt-injection defence rate than GPT-5.6 Terra.
OpenAI released GPT-6.1 Terra and said it was its latest mid-range model. The company said API pricing had been cut to half the level of GPT-5.6 Terra, with input tokens at $1 per million and output tokens at $5.
OpenAI said Terra was built for office work and automated software control. It said the model was available now to Plus, Pro, Enterprise and Edu users in ChatGPT Work and Codex, as well as through the API.
Artificial Analysis said Terra offered a better result on intelligence per dollar than GPT-5.6 Terra. It said the model gained 6 points on the Intelligence Index at maximum reasoning effort, while cost per task fell from $1.40 to $0.40; at lower effort levels, it said the gains were 7 to 13 points. On Coding Agent Index v2.0, Terra scored 48, compared with 54 for GPT-6.1 Sol and 31 for GPT-6 Luna.
OpenAI said Terra was especially strong at computer control. On Agents' Last Exam it matched GPT-6 Astra at one-tenth of the token price, and the article said it finished more than three times faster than GPT-6 Astra and GPT-6.1 Sol. OpenAI also said Terra met its "Critical" cybersecurity threshold and inherited GPT-6 Astra’s safeguards; the system card put its indirect prompt-injection defence rate at 98.8 percent, up from 96.2 percent for GPT-5.6 Terra.
Named in this story
Companies
- OpenAI
- released GPT-6.1 Terra and set its API pricing
- Artificial Analysis
- provided benchmark comparisons and price-per-task figures
- Anthropic
- owned the Claude models used for comparison
Products and systems
- GPT-6.1 Terra
- the new mid-range model at the centre of the story
- GPT-5.6 Terra
- the predecessor used for price and performance comparisons
- ChatGPT Work
- one of the places where Terra became available
- Codex
- another product through which Terra became available
- Claude Sonnet 5.5
- the Anthropic model used in the pricing comparison
- Claude Opus 5.5
- the Anthropic model used in token-efficiency comparisons
- GPT-6.1 Sol
- one of OpenAI’s higher-tier models used as a benchmark reference
- GPT-6 Luna
- the lower-scoring OpenAI model in the Coding Agent Index comparison
- GPT-6 Astra
- a higher-tier OpenAI model used as a comparison point for control and safety tests
How the source tells it
The piece is written as a product launch story, with promotional benchmark comparisons and OpenAI’s own claims carrying most of the weight.
- hype headline-level superlatives and repeated 'new benchmark' framing around price and performance
- vendor boosterism OpenAI’s claims about pricing, capability and safety are carried through the article with little independent challenge