OpenAI Slashes GPT-5.6 Luna Prices by 80%, Terra by 20%

OpenAI has reduced the API pricing of its GPT-5.6 Luna and Terra models, lowering Luna’s price by 80% and Terra’s by 20%, while introducing a new Fast mode for GPT-5.6 Sol that delivers up to 2.5 times faster responses than the standard mode.

The changes take effect immediately. GPT-5.6 Luna will now cost $0.20 per million input tokens and $1.20 per million output tokens, while GPT-5.6 Terra is priced at $2 per million input tokens and $12 per million output tokens. Sol pricing remains unchanged.

The company said the lower prices are also reflected in how usage is counted against paid subscriptions in ChatGPT Work and Codex, allowing customers to consume fewer credits for Terra and Luna workloads.

“Making advanced intelligence more abundant and affordable is central to OpenAI’s mission to ensure AGI benefits all of humanity,” the company said. “These changes put that commitment into practice.”

OpenAI also replaced its Priority Processing offering with Fast mode in the API. For GPT-5.6 Sol, Fast mode provides up to 2.5 times faster performance than Standard processing at twice the price, with no change in model intelligence. Existing API requests tagged as priority will automatically use Fast mode.

The company said the pricing changes are intended to help businesses match model capability with the requirements of different tasks. It said Luna can be used for high-volume, multi-step workflows, while Sol can be reserved for tasks requiring higher reasoning capabilities.

According to OpenAI, GPT-5.6 Luna delivers performance comparable to models that were considered frontier-class a year ago at “roughly 6 cents on the dollar per task” and at nearly nine times the speed. The company also said Luna outperformed Claude Fable 5 on the Agents’ Last Exam benchmark at an estimated cost per task nearly 99% lower.

OpenAI attributed the lower prices to improvements across its models, inference systems and what it calls an ‘agentic harness’ that connects models to tools and context. The company said that better routing, more efficient production software, and improved context management reduce the compute, time, and tokens needed to complete tasks.

The company also disclosed that GPT-5.6 Sol is helping optimise its own infrastructure. Within what OpenAI described as a human-led process, the model rewrote production kernels, designed and ran experiments to improve token generation, and monitored training runs.

“The kernel work helped reduce the end-to-end cost of serving the model by 20%, while its experiments increased token-generation efficiency by more than 15%,” OpenAI added.

It further said that it continues to expand its compute infrastructure while matching workloads to the hardware best suited to run them. This strategy allows enterprises to use lower-cost models such as Luna and Terra for routine, large-scale tasks, while using Sol for workloads where faster response times justify the higher cost.

GPT-5.6 Terra and Luna remain available through ChatGPT Work, Codex and the OpenAI API. Free and Go users can access Terra in ChatGPT Work and Codex, while Plus, Pro, Business and Enterprise subscribers can use both Terra and Luna. The new API pricing took effect on July 30, with AWS rollout beginning later the same day.