OpenAI has once again driven up the price-performance of large models.
Just now, OpenAI officially released its new-generation model, GPT-6.1 Terra, which focuses on faster response times, stronger visual understanding and computer-operation capabilities, and lower usage costs.
Compared with the previous-generation GPT-5.6 Terra, the new model’s standard API price is cut by 50%, yet it shows clear improvements in complex task execution, factual accuracy, and coding. In some professional task tests in particular, GPT-6.1 Terra can already approach the performance of the flagship GPT-6 Astra, at only one-tenth of its token price.
OpenAI positions it as an efficient model for high-frequency work scenarios, hoping to let more businesses and developers put AI agents to real use in everyday work at lower cost.
