OpenAI adds GPT-6 Luna and Sol, touts lower prices

OpenAI launched GPT-6 Sol and GPT-6 Luna in a fast follow to GPT-6 Astra. As OpenAI rounded out its GPT-6 family it also signaled that prices would come down.

In a nutshell, GPT-6 Sol and GPT-6 Luna are 50% less expensive than GPT-5.6. OpenAI CEO Sam Altman said in a thread on X that the company wants to "feature the best model at every price point and be the best at every modality."

The news from OpenAI landed shortly after Anthropic outlined Opus 5.5, which will be 40% less than Opus 5.

Simply put, it's pretty clear that frontier LLM model prices are coming down. After all, OpenAI and Anthropic are competing against good enough open models in the enterprise and the likes of Meta and SpaceXAI, which are driving prices lower. Don't be surprised if Meta aims to undercut both OpenAI and Anthropic on pricing during its Connect conference.

For now, it's swell that the two halves of the frontier AI duopoly are competing on price. Soon, it'll be Google's serve with Gemini 4.

Key details about OpenAI's GPT-6 Sol and GPT-6 Luna include:

• API prices for Sol and Luna are 50% cheaper than GPT-5.6 promotional pricing.
• Luna and Sol were trained on similar methods as Astra.
• OpenAI led with price in its blog and then focused on performance.
• The company is looking to combine price and performance and cost per task.
• GPT-6 Sol makes half as many mistakes as its predecessor. GPT-6 Luna matches GPT-5.6 Sol at higher effort levels at a much lower cost.
• OpenAI also said Astra's improved communication style will come to Sol and Luna.

GPT‑6 Sol and GPT‑6 Luna are available in ChatGPT Work and Codex now for Plus, Pro, Business, Enterprise and Edu plans. Free and Go users can access GPT‑6 Luna in the desktop app.

The backdrop:

• Nvidia’s Nemotron hit enterprise AI inflection point
• Why you should arbitrage your AI models like your vendors do
• Nvidia’s $12.93 billion Hugging Face acquisition validates open weight models
• Open-Weight Models Are Gaining Ground in Enterprise AI
• Get ready for US open LLMs and just in time
• Rightsizing open models may cut your AI inference spend
• The most important LLM benchmark today: Price