OpenAI has slashed the price of GPT-5.6 Luna, its fastest and most affordable model, by 80 percent, dropping the cost to 20 cents per million input tokens and $1.20 per million output tokens. The balanced Terra tier fell 20 percent to $2 and $12, while flagship Sol pricing is unchanged at $5 and $30.
The cuts, effective July 30, arrive three weeks after the GPT-5.6 family launched and reflect what OpenAI describes as efficiency gains across models, inference systems, and agent harnesses. The company says Sol autonomously rewrote production kernels, cutting serving costs by 20 percent and lifting token-generation efficiency by more than 15 percent.
OpenAI claims Luna now undercuts DeepSeek on input pricing and delivers frontier-class performance at a fraction of prior costs, beating Claude Fable 5 on the Agents’ Last Exam benchmark at an estimated cost per task nearly 99 percent lower. The company also introduced Fast mode, which serves GPT-5.6 Sol up to 2.5 times faster at double the price.
Early customers are leaning into the economics. Replit president Michele Catasta called Luna the closest we’ve come to intelligence too cheap to meter, and Cognition says the model now handles routine coding work inside Devin Fusion. The price changes began rolling out to AWS later on July 30.
