According to OpenAI, the company has previewed GPT-5.6 Sol, described as a next-generation model that demonstrates stronger capabilities across multiple technical domains. The model specifically shows improvements in coding, science, and cybersecurity applications, according to the announcement.
OpenAI states that GPT-5.6 Sol is paired with what the company describes as its most advanced safety stack to date. While the announcement confirms the model’s enhanced technical capabilities and safety features, specific details about performance benchmarks, availability timelines, or the technical architecture of the safety systems were not provided in the source material.
The preview represents OpenAI’s continued development of more capable AI models with an emphasis on both performance improvements and safety measures. The focus on coding, science, and cybersecurity suggests the model is being positioned for technical and specialized applications.