Earlier this month, OpenAI launched GPT-6 Astra, which it described as its most powerful and capable model yet. The company said Astra was the “world’s best model” for activities including computer work and coding. OpenAI is now expanding the GPT-6 generation with updated versions of its smaller Sol and Luna models.
“GPT-6 Astra introduced a new generation of intelligence; these models extend its benefits by making that intelligence more efficient and accessible,” OpenAI wrote. The first Sol and Luna models were introduced earlier this year as different tiers in OpenAI’s model hierarchy. Sol is designed for complex tasks such as coding, while Luna is intended for clerical work, including summarizing documents, extracting information, and answering quick questions.
OpenAI is emphasizing improvements in efficiency and affordability, including a substantial price reduction for API access . The GPT-6 Sol and Luna models will be available at half the cost of the GPT-5.6 series. OpenAI attributes the reduction to improvements in caching and inference .
OpenAI says the new models produce more factually accurate answers and have a lower error rate in coding tasks. In its internal factuality evaluation , based on de-identified real-world conversations where users flagged model mistakes, OpenAI said GPT-6 Sol makes about half as many mistakes as its predecessor. The company said this gives Sol Astra-level reliability at a much lower cost.
OpenAI claimed that GPT-6 Sol and Luna handle tasks substantially better than Anthropic’s top models, including Fable and Opus. Anthropic released a new version of Opus 5.5 just 90 minutes before OpenAI’s announcement. The new Sol and Luna versions are available in ChatGPT Work and Codex for most paid accounts and through the ChatGPT API, while Luna is also available in the desktop app and for Free and Go users.