OpenAI has cut GPT-5.6 Luna pricing by 80% and GPT-5.6 Terra pricing by 20%, while introducing a premium Fast mode for GPT-5.6 Sol that offers up to 2.5 times the standard speed.
OpenAI has reduced API prices for its GPT-5.6 Luna and Terra models and introduced a faster, higher-priced serving option for GPT-5.6 Sol.
In its announcement, OpenAI said GPT-5.6 Luna is now 80% cheaper and GPT-5.6 Terra is 20% cheaper. The company attributed the changes to advances in serving efficiency, which can reduce the cost of operating models for API customers.
According to Axios, Luna now costs $0.20 per million input tokens and $1.20 per million output tokens. Terra is priced at $2 per million input tokens and $12 per million output tokens.
The reductions widen the pricing gap between Luna, OpenAI's lower-cost option, and Terra, which occupies the middle of the GPT-5.6 range. OpenAI's GPT-5.6 launch material describes Sol, Terra, and Luna as a three-tier model family.
Alongside the price reductions, OpenAI added an API Fast mode for GPT-5.6 Sol. The company said the option can provide up to 2.5 times the speed of Sol's Standard mode at twice the price.
OpenAI said Fast mode does not alter the model's intelligence. Instead, it is a serving choice for developers that place a higher value on response time and are prepared to pay more for lower latency.
That distinction matters for teams choosing models for different workloads. A high-volume application whose main constraint is token cost may find Luna's lower rates more relevant. Developers seeking a more capable middle-tier option may look to Terra's revised price. Applications such as interactive assistants or real-time workflows, where delayed responses can be more costly than higher token charges, may be candidates for Sol Fast.
The update illustrates how AI API providers are competing on operational characteristics as well as model capability. Published token rates affect the economics of large-scale deployments, while speed options allow customers to trade more spending for faster responses.
For GPT-5.6 users, the immediate changes are straightforward: Luna and Terra have lower listed input and output token prices, while Sol has an additional premium mode designed for speed-sensitive workloads. OpenAI has presented the Fast option as an inference-performance configuration rather than a separate model release or an intelligence upgrade.
The company did not characterize the reductions as temporary. Its announcement frames them as part of an effort to improve the price-performance balance available through the GPT-5.6 API family.
Lower prices for two GPT 5.6 tiers OpenAI has reduced API prices for its GPT 5.6 Luna and Terra models and introduced a faster, higher priced serving option for GPT 5.6 Sol.
In its announcement, OpenAI said GPT 5.6 Luna is now 80% cheaper and GPT 5.6 Terra is 20% cheaper.
The company attributed the changes to advances in serving efficiency, which can reduce the cost of operating models for API customers.
Continue reading
