OpenAI rolls out budget-priced mid-tier GPT-6.1 Terra
OpenAI has added a mid-range model to its GPT-6.1 family, offering strong performance at a lower price.
The AI business' general-purpose models span four tiers. Products named "Sol" or "Astra" are its high-end offerings and come with price tags to match. "Luna" models are its cheapest offerings. This new model gets the "Terra" treatment the company reserves for mid-range models.
"Terra continues to advance the cost-intelligence frontier across our model families, providing a new balance of speed, capability, and cost," OpenAI said. "Its standard API prices are 50 percent lower than those of GPT-5.6 Terra and GPT-6.1 Sol, making it practical for developers to automate high-volume work, like reading incoming invoices and entering their details into accounting software."
Artificial Analysis, a site that tracks model performance and cost, notes that GPT-6.1 Terra pushes the cost efficiency frontier. "It is the first Terra-class model since GPT-5.6, and is more intelligent for less than a third of the cost," the site says.
GPT-6 Luna ($0.10/M input tokens, $0.50/M output) occupies the lowest tier of OpenAI's model lineup in terms of cost and capability, followed by GPT-6.1 Terra ($1/M, $5/M), GPT-6.1 Sol ($2/M, $10/M), and GPT-6 Astra ($10/M, $50/M).
In a comment posted to Hacker News, Ted Sanders, a member of technical staff at OpenAI, framed the differences between the various GPT-6.1 models thus: "In the case of GPT-6.1 Terra, I'd say Terra is probably better than Astra and Sol at technical communication, similar on vision and computer use (but much faster!), and a bit worse on most other things. For API developers, it's a nice midrange option if you want something cheaper than Sol but a lot smarter than Luna (especially for vision/computer use/communication)."
"For subscribers, it's not a bombshell release - if you're hitting usage limits, maybe it helps you go a bit further, and if you want clearer communication or faster computer use, it can be a nice option to select."
Terra's knowledge cutoff date, June 1, 2026, means it is the most up-to-date of its model family. Perhaps that counts for something if you're asking questions about events after April 30, 2026 - the training data cutoff for Astra and Sol.
In terms of model safety, Terra handles harmful requests a bit less frequently than GPT-6.1 Sol but also avoids unnecessary refusals more often than its more capable sibling.
But as noted by Max Woolf, a data scientist, in a recent blog post, the real story here is the economics of AI models. Despite the widespread expectation that OpenAI and Anthropic will have to raise the cost of AI inference (or reduce the value provided via subscription) to bring revenue closer to the extravagant sums these companies are spending on model training, inference, and infrastructure, he notes that prices keep coming down and model performance keeps going up.
"The idea that token prices will suddenly rise is more a fantasy that is not supported by both behavioral economics and the actual trends of LLM prices going down significantly over the last few months for better models, and not just from OpenAI and Anthropic," he observes. "As long as scalers can price at a cost-per-token lower than the revenue-per-token, then those inference providers will reduce token costs to compete with each other to a point that OpenAI/Anthropic will have to lower their prices to compete too."
As to whether these price trends suggest that AI companies have a sustainable business, investors may soon be able to make that bet on the Nasdaq exchange after Anthropic floats. (R)