Artificial intelligence pioneer OpenAI has expanded its advanced GPT-6 series by introducing two new AI models named GPT-6 Sol and GPT-6 Luna. These models are engineered to bring the advanced capabilities introduced in the recently revealed GPT-6 Astra into more cost-effective options for developers and users alike.
Key Features of GPT-6 Sol and GPT-6 Luna
The two models cater to distinct operational needs. According to OpenAI, GPT-6 Sol is specifically tailored for rigorous reasoning, advanced coding, and complex agentic workflows. In contrast, GPT-6 Luna focuses on rapid, high-volume everyday tasks where speed and processing economy matter the most.
Both models deliver notable improvements over the preceding GPT-5.6 generation, offering better factuality, enhanced coding proficiency, and superior communication skills. Through the API, both variants support multimodal inputs including text and images. They boast a massive 1.05 million token context window alongside a maximum capacity of 128,000 output tokens.
Competitive API Pricing Structure
To make large-scale workloads more viable, OpenAI has priced the new models significantly lower than the GPT-5.6 generation. GPT-6 Sol is priced at 2 dollars for 1 million input tokens and 10 dollars for output tokens. Meanwhile, GPT-6 Luna offers extreme affordability at 0.10 dollars per 1 million input tokens and 0.50 dollars for output tokens.
These rates mean developers running heavy AI workloads will benefit from substantial cost savings. OpenAI notes that the pricing for Sol and Luna is roughly 50 percent lower than the initial promotional rates of the previous GPT-5.6 series.
Platform Availability and Rollout
GPT-6 Sol and GPT-6 Luna are accessible via the API right away. Following the announcement, both models have rolled out across ChatGPT Work and Codex for Plus, Pro, Business, Enterprise, and Edu users. Free and Go tier users are given the option to utilize GPT-6 Luna directly within the desktop application.
Availability may still vary depending on the specific product interface. Within the broader GPT-6 family, OpenAI positions Astra as the most capable overall model, Sol as a balanced option for demanding tasks, and Luna as the most efficient model for high-frequency operations.
