OpenAI has announced price reductions for its API models "GPT-5.6 Luna" and "GPT-5.6 Terra," as well as performance speed-ups for "GPT-5.6 Sol."
Pricing & LimitsOpenAIgpt-5.6-lunaGPT-5.6 TerraGPT-5.6 Sol
OpenAI Slashes GPT-5.6 Luna Price by 80%, Introduces Fast Mode for API Acceleration
This article is a translation. Read the Japanese original
Specifically, the lightweight Luna model will see an 80% price reduction, while the balanced Terra model will be reduced by 20%. According to OpenAI, Luna delivers performance comparable to frontier-class models from one year ago, while reducing per-task costs by approximately 99% and increasing speed by about nine times. These price cuts will also be reflected in the usage calculations for paid subscriptions of ChatGPT Work and Codex.
Additionally, a new "Fast mode" is being introduced to the API. Replacing the previous Priority Processing, this mode enables GPT-5.6 Sol to achieve up to 2.5 times the speed of Standard processing at twice the cost, while maintaining response performance. The price tags are $0.20 per million input tokens and $1.20 per million output tokens for Luna, and $2 per million input tokens and $12 per million output tokens for Terra.
Sources
- 「AI利用料は90％下落、でも支出増」の怪――AIコスト問題で注目の「ジェボンズのパラドックス」から考える （ITmedia AI+, 2026-09-11）
- OpenAI Official Blog