OpenAI

OpenAI has released GPT-6 Sol and GPT-6 Luna, two new API models that sit below the flagship GPT-6 Astra in the same family. Astra launched earlier in the month as the company's most powerful and capable model, positioned for computer work and coding , and Sol and Luna were developed using methods similar to Astra's for improvements in professional work, factuality, coding, and computer use . For developers, the pitch is that the same training approach now reaches cheaper tiers. The launch matters mostly on price. Both models cost 50% less than the GPT-5.6 promotional pricing they replace, with OpenAI attributing the reduction to caching and inference improvements it is passing through to API customers . Independently reported figures put Sol at $2/$10 per million input/output tokens and Luna at $0.10/$0.50 , which frames the tagline: Sol is aimed at building and iterating on serious workloads, while Luna is priced for high-volume production traffic. The release lands just nineteen days after GPT-6 Astra and ahead of OpenAI's September 29 DevDay event , giving teams a full tiered lineup to plan around before the conference. Distribution extends beyond the API. Sol and Luna are rolling out in ChatGPT Work and Codex for Plus, Pro, Business, Enterprise, and Edu users, while Free and Go users get access to Luna in the desktop app . For founders and operators already routing traffic through OpenAI, the practical question this launch forces is whether to move existing 5.6 workloads onto Luna for cost, keep Sol for reasoning-heavy paths, and reserve Astra for the hardest tasks.