OpenAI Slashes GPT-5.6 API Prices: Luna Down 80%, Terra Cut 20%
The move makes high-volume AI workflows economically viable for enterprises while introducing a premium Fast mode for the flagship Sol tier.
OpenAI cut API pricing for two GPT-5.6 tiers effective July 30, 2026, making enterprise AI deployment significantly more affordable while adding a premium Fast mode for latency-critical workloads.
GPT-5.6 Luna, the fastest and most cost-efficient tier, now costs $0.20 per million input tokens and $1.20 per million output tokens—an 80% reduction. The balanced GPT-5.6 Terra tier dropped 20% to $2 per million input tokens and $12 per million output tokens. The changes are rolling out to AWS starting July 30.
Fast Mode for Sol
Alongside the price cuts, OpenAI introduced "Fast mode" for its flagship GPT-5.6 Sol model. The option delivers up to 2.5x faster processing speeds than standard Sol at twice the price, giving developers a latency-sensitive option for time-critical applications.
Benchmark Claims
OpenAI states that Luna outperforms Fable 5 on professional workloads measured by its "Agents' Last Exam" benchmark, at an estimated cost per task nearly 99% lower. The claim comes from OpenAI's own announcement and has not been independently verified by third-party evaluators.
Why the Cuts Matter
The pricing strategy targets high-volume enterprise workflows such as large-scale document analysis and routine code implementation. By making Luna economically viable for bulk processing, businesses can route complex tasks to Sol while using Luna for lower-complexity, high-volume work—reducing overall AI deployment costs at scale.
The three-tier GPT-5.6 family (Sol, Terra, Luna) follows OpenAI's efforts to use its own models to optimize production kernels and token generation efficiency.
"Making advanced intelligence more abundant and affordable is central to OpenAI's mission to ensure AGI benefits all of humanity," the company said in its announcement.
Unconfirmed Checkpoints
Secondary reports have mentioned two unlabeled OpenAI checkpoints, Zinc and Magnesium, appearing on DesignArena, though OpenAI has not officially confirmed these model names. The core pricing changes, however, are confirmed across multiple sources including OpenAI's official blog and tech industry outlets.
The July 30 price cuts position OpenAI to compete more aggressively on cost-sensitive enterprise contracts while maintaining premium tiers for latency-critical and complex reasoning workloads.