OpenAI Launches GPT-6 Sol and Luna, Slashing API Inference Costs by 50%

OpenAI has officially launched GPT-6 Sol and GPT-6 Luna, expanding its next-generation model family alongside flagship GPT-6 Astra. The release is aimed at shifting the frontier of cost efficiency, delivering Astra-grade reasoning, coding, computer-use, and factual accuracy to high-throughput and cost-sensitive workloads.
Thanks to architectural advancements in prompt caching and inference optimization, OpenAI cut API pricing by 50% compared to previous GPT-5.6 promotional rates:
- GPT-6 Sol: Positioned as the primary workhorse model for complex professional automation and coding, priced at $2.00 per million input tokens and $10.00 per million output tokens.
- GPT-6 Luna: Engineered for ultra-fast, low-latency tasks and agentic loops at $0.10 per million input tokens and $0.50 per million output tokens.
Both models are live today across the API, ChatGPT Work, Codex, and GitHub Copilot for Plus, Pro, Business, Enterprise, and Edu tiers, with Free and Go users gaining desktop access to Luna. The launch comes as frontier AI labs face mounting pressure to balance massive infrastructure compute costs with scalable, unit-economic enterprise adoption.