OpenAI released two new GPT-6 variants, GPT-6 Sol and GPT-6 Luna, with substantial price cuts and infrastructure upgrades aimed at lowering costs for developers and businesses. GPT-6 Sol is priced at $2 per million input tokens and $10 per million output tokens, down from GPT-5.6 Sol’s $4 and $20. GPT-6 Luna drops to $0.10 and $0.50 from $0.20 and $1.20. The company reports reduced hallucination and coding-deception rates: internal figures cite 1.3% for Sol and 2.8% for Luna versus GPT-5.6 Sol’s 10.4%.
Alongside token pricing, OpenAI introduced prompt-caching upgrades that raise default cache hit rates, add diagnostics and explicit breakpoints, and let developers toggle reasoning effort and swap tools without invalidating cached context. These changes can cut cached input token costs by up to 90% and reduce latency for long conversational contexts, materially lowering the effective expense of sustained agentic pipelines and multi-step workflows.
OpenAI positions Astra as the flagship for the hardest tasks while Sol and Luna are faster, cheaper options for everyday work. In OpenAI’s own benchmarks, GPT-6 Sol scored 33.2% on AutomationBench at $0.27 per task and 56.4% on Agents’ Last Exam at its highest effort setting, with cost and performance comparisons cited versus Anthropic’s Claude Opus 5 and Claude Fable 5.1. The new models are available in ChatGPT Work and Codex for Plus, Pro, Business, Enterprise, and Edu subscribers, with Luna also reaching some free and Go desktop users; rollout is being staged to preserve service stability.
This summary is composed by the cFlash AI editor from multiple public sources, under human supervision. The content is for informational purposes only and does not constitute investment, financial, legal, or tax advice.