OpenAI has introduced two new models, GPT-6 Sol and Luna, which offer a 50 percent price reduction compared to the previous GPT-5.6 versions. The cost for GPT-6 Sol is set at $2 per million input tokens and $10 per million output tokens. Luna is priced at $0.10 for input and $0.50 for output.
The company attributes these lower costs to improvements in inference and caching. OpenAI has also updated its prompt caching for GPT-6, providing a 90 percent discount on cached input tokens. Developers now have access to a new dashboard and diagnostic tools to manage cache usage and adjust reasoning effort without invalidating the cache.
GPT-6 Sol is designed for complex tasks such as coding and data analysis, while Luna focuses on high-volume, well-defined tasks like document summarization. The model Terra has been removed from the lineup. OpenAI positions these new models as cost-effective alternatives to Anthropic's Claude series, citing lower costs per task on benchmarks like AutomationBench.
Independent analysis from Artificial Analysis indicates that while per-task costs have been halved, intelligence scores remain largely at GPT-5.6 levels. The report notes performance gains in some areas and regressions in others, including a drop in Elo points on the GDPval-AA v2.1 benchmark, which the firm attributed to lower presentation quality.
