According to OpenAI's developer documentation, the company on November 20 introduced completion windows for Responses API background mode, offering token rate discounts of 30% for 2-hour and 40% for 8-hour windows compared to standard rates. The feature is currently in public testing across GPT-6.2 Astra, GPT-6.2 Sol, GPT-6.1 Terra, and GPT-6 Luna.
Discounted windows require Standard processing and do not support streaming output. Requests using the same continuation_id receive higher queue priority. Beginning with GPT-6.2 Astra and Sol, the company is discontinuing Flex processing for new models.