OpenAI announced price reductions and performance improvements for the GPT-5.6 model family, passing efficiency gains to customers. GPT-5.6 Luna, the fastest and most affordable model, costs 80% less, while GPT-5.6 Terra, the balanced model, costs 20% less. These price changes also reduce how usage counts against paid subscriptions in Codex and ChatGPT Work. The company introduced Fast mode in the API for GPT-5.6 Sol, delivering up to 2.5× faster speeds than Standard processing at twice the price, with no intelligence change. Fast mode replaces Priority Processing and is backward compatible.

The post frames efficient AI use as matching model choice to the outcome, cost of error, urgency, and scale. Luna delivers performance comparable to frontier-class models from a year ago at roughly 6 cents on the dollar per task and nearly nine times the speed. On professional work measured by Agents’ Last Exam, Luna outperforms Fable 5 at an estimated cost per task nearly 99% lower. A coding workflow example uses Sol to resolve uncertainty and define plans, then Luna to implement changes, run tests, and evaluate results.

The efficiency gains come from improvements across model architecture, inference systems, and the agentic harness. GPT-5.6 Sol itself helped find gains: within a human-led process, Sol autonomously rewrote and optimized production kernels, designed and ran hundreds of experiments to improve token generation, and monitored training. The kernel work reduced end-to-end serving cost by 20%, and experiments increased token-generation efficiency by more than 15%.

As of July 30, API pricing is $2 per million input tokens and $12 per million output tokens for Terra, and $0.20 per million input tokens and $1.20 per million output tokens for Luna. Sol pricing remains unchanged. Pricing changes begin rolling out in AWS later. ChatGPT and Codex subscription prices and quota budgets remain unchanged, with Terra and Luna usage now consuming fewer credits.