OpenAI Expands GPT-6 Line With Cheaper Sol and Luna Models

OpenAI has added updated versions of its smaller Sol and Luna models to the GPT-6 generation, following this month's launch of GPT-6 Astra, which the company called its most powerful model yet and the "world's best model" for tasks including computer work and coding.

Sol and Luna first arrived earlier this year as different tiers in OpenAI's lineup: Sol for complex work like coding, and Luna for high-volume tasks with a clear goal, such as summarizing documents, extracting information, or answering quick questions.

OpenAI says the new versions are more efficient and affordable, with API access at half the cost of the 5.6 series of Sol and Luna, which it credits to improvements in caching and inference. The company also claims better factual accuracy - on its internal factuality evaluation, based on de-identified real-world conversations where users flagged mistakes, GPT-6 Sol makes about half as many mistakes as its predecessor, reaching Astra-level reliability at much lower cost - and a lower error rate in coding.

As usual, OpenAI claims its models outperform Anthropic's top offerings, such as Fable and Opus. Anthropic released a new version of Opus 5.5 just 90 minutes before OpenAI's release.

The new Sol and Luna versions are available now in ChatGPT Work and Codex for most paid accounts and in the ChatGPT API. Luna will also be available in the desktop app and for Free and Go users, with a gradual rollout across ChatGPT's app and website throughout the day.

Source: TechCrunch