Washington, October 01 (QNA) – OpenAI has launched "GPT-6.1 Sol", a model designed for programming and complex tasks.

In a statement, the company said that the new model reduced the error rate in real-world accuracy tests—when utilizing low-level reasoning—to 7.7 percent, down from 11.4 percent in the "GPT-6 Sol" model, with performance approaching that of "GPT-6 Astra".

The model enables the execution of programming tasks using AI agents—including task planning, execution, and tool usage—rather than simply suggesting code snippets.

OpenAI noted that the model is clearer in regards to its limitations, and adheres more strictly to user constraints and safety guidelines compared to the previous version.

GPT-6.1 Sol is now available to Plus, Pro, Business, Enterprise, and Edu users within ChatGPT Work and Codex; however, it has not yet been added to the standard consumer chat model selection menu across all ChatGPT interfaces. (QNA)