OpenAI has upgraded its medium-sized Terra model from GPT-5.6 to GPT-6.1, touting improved speed, computer use capabilities, and lower pricing. The update follows last month's release of the GPT-6 Sol and Luna models, which initially excluded Terra.
GPT-6.1 Terra balances intelligence and affordability as part of OpenAI's four model classes: Luna, Terra, Sol, and Astra. The new model excels at communication and computer-based tasks like drafting proposals, interpreting technical drawings, filling out forms, and editing videos.
Pricing for GPT-6.1 Terra is 50% lower than the replaced GPT-5.6 Terra version, making it ideal for automating computer work tasks. On internal benchmarks, the error rate at low reasoning effort dropped from 53.6% to 16.7%, while maximum effort error rates fell 18 percentage points to 4.3%, nearly matching GPT-6 Astra and GPT-6.1 Sol.
The GPT-6.1 Terra model is available today through ChatGPT Work, Codex, and the API.
Source: 9to5Mac · Summarized by HeadlinesBriefing