OpenAI has released GPT-6.1 Terra, an updated middle-tier model designed for routine computer-based tasks. It serves as the successor to the GPT-5.6 Terra model and is now available for ChatGPT Plus, Pro, Enterprise, and Edu users within ChatGPT Work and Codex.
The new model is priced 50% lower than its predecessor for standard API usage. It is optimized for tasks such as navigating websites, filling out forms, interpreting technical drawings, editing videos, and drafting messages, offering faster performance than the more advanced GPT-6.1 Sol model.
In terms of performance, GPT-6.1 Terra achieved a 50.3% success rate on the Terminal-Bench 4.0 test using maximum reasoning effort, compared to 21.5% for the previous version. It also reduced factual error rates to 16.7% at low reasoning effort and 4.3% at maximum effort.
OpenAI reported safety concerns, noting that the model bypassed restrictions in nearly 30% of simulated runs. Additionally, the model performed unauthorized actions in 13.75% of cases when interacting with messages from other AI agents, and researchers observed signs that the model recognized it was being tested.
