OpenAI launches GPT-6 Sol and Luna, boasting lower cost and fewer mistakes
TechCrunch Lucas Ropek ● Covered by 2 sources
OpenAI updated its Sol and Luna AI models. They cost less now, and OpenAI says they make fewer mistakes.
Based on reporting by TechCrunch, Lucas Ropek — read the original for the full story.
Summary, retelling and take written by AI under human oversight; images are AI-generated illustrations. How we work · Report an error
OpenAI has pushed another round of GPT-6 models into the market, this time updating Sol and Luna after launching GPT-6 Astra earlier this month. Astra was billed as the company’s most powerful model yet, but OpenAI is now focusing on making the smaller models cheaper and easier to use at scale.
Sol and Luna sit lower in OpenAI’s model lineup. Sol is aimed at tougher work such as coding, while Luna is pitched for clerical jobs: summarizing documents, pulling out information, and answering quick questions. OpenAI says these updated versions are meant to spread the same intelligence more efficiently.
The biggest change is price. OpenAI says the GPT-6 series will be available at half the cost of the 5.6 series for Sol and Luna, and it credits caching and inference improvements for that drop. The company is also making a reliability claim, saying the new models are better at factual answers and produce fewer coding errors.
OpenAI pointed to an internal factuality evaluation based on de-identified real-world conversations where users had flagged mistakes. In that test, the company says GPT-6 Sol makes about half as many mistakes as its predecessor and reaches Astra-level reliability at much lower cost.
The release also keeps the rivalry with Anthropic front and center. OpenAI says Sol and Luna outperform Anthropic’s top models, including Fable and Opus. Anthropic had just released a new version of Opus 5.5 90 minutes before OpenAI’s announcement, which tells you everything you need to know about how tightly wound this race is right now.
My take — AI-written commentary, not fact-checked reporting
This is the part of the AI race that actually matters: not the loudest model, but the one people can afford to run without holding their breath. Cheaper models with fewer mistakes are the real product, even if the press release still wants a trophy. The industry keeps pretending bigger is always better; the bill says otherwise.
Read more about this at: TechCrunch
