OpenAI Releases GPT-6.1 Terra, API Pricing Reduced by 50% Compared to Previous Generation

OpenAI announced the release of GPT-6.1 Terra, focusing on enhancements to visual reasoning, computer operation, factual accuracy, and software engineering capabilities. The model is designed for daily office workflows and high-frequency automation tasks, capable of processing technical drawings, completing forms, editing videos, and developing software. In terms of performance, OpenAI reported that GPT-6.1 Terra reduced its factual error rate in internal complex prompt tests from 53.6% (observed in GPT-5.6 Terra at low reasoning intensity) to 16.7%, with the error rate further dropping to 4.3% under highest reasoning intensity. In the Terminal-Bench 4.0 programming benchmark, the model achieved a task completion rate of 50.3% at peak reasoning intensity, surpassing GPT-5.6 Terra's 21.5%.