OpenAI introduced the GPT-6 Sol and Luna models

OpenAI has released two new language models — GPT-6 Sol and Luna, aimed at increasing performance while reducing operating costs. The new models show a result of 1347 points in comprehensive testing, which took 11 hours of continuous operation.

The key advantage of the update was a reduction in API costs by 50% compared to the previous version of GPT-5.6. The developers focused on optimizing the architecture for professional tasks, which allowed for more accurate code generation and increased factual accuracy of responses.

The Sol and Luna models demonstrate improved performance in logical reasoning and long context handling. According to the developer community, the transition to the new architecture allows for a reduction in response time while maintaining high-quality generation in complex usage scenarios.

The update is aimed at integration into corporate systems, where the cost of token processing is critical. The twofold price reduction makes the implementation of agents based on GPT-6 more accessible for scalable applications that require high accuracy in programming and data analytics.