OpenAI has introduced GPT-6.1 Sol, a new artificial intelligence model designed to bring advanced reasoning and agentic capabilities to a wider range of professional workflows at a lower cost.
Presented during the company's DevDay event, GPT-6.1 Sol is positioned as a more efficient option for tasks including software development, debugging, document analysis and multi-step computer workflows. OpenAI says the model approaches the performance of GPT-6 Astra in several demanding professional use cases while reducing standard input and output token pricing to roughly one-fifth.
Improved accuracy and workflow reliability
According to OpenAI, GPT-6.1 Sol builds on GPT-6 Sol with stronger results in complex programming tasks and broader document comprehension. The model is also designed to handle sequential workflows with greater consistency, supporting teams that use AI for structured business and technical processes.
The company reports gains in factual accuracy, particularly when the model operates with lower reasoning effort. In its internal evaluations, the rate of responses containing factual errors declined from 11.4% in GPT-6 Sol to 7.7% in GPT-6.1 Sol under this setting.
OpenAI also says the model is more transparent about its limitations and better at following user-defined instructions and operational constraints. These refinements aim to make AI-supported work more dependable across coding, research and enterprise productivity environments.
Availability
GPT-6.1 Sol is now available for Plus, Pro, Business, Enterprise and Edu users through ChatGPT Work and Codex. The model is not currently offered in Chat.
By combining near-frontier performance with lower operating costs, GPT-6.1 Sol could help make sophisticated AI tools a more practical part of everyday professional creation and problem-solving.
