OpenAI is preparing to launch a new “router” feature for ChatGPT, which will automatically select the best AI model for each user request. Currently, ChatGPT Plus subscribers and higher-tier plans have access to seven different models, including GPT-4o, o3, o4-mini, o4-mini-high, GPT-4.5 (Research Preview), GPT-4.1, and GPT-4.1-mini. Each of these models has its strengths and is used for different tasks: some work better with math and programming, while others excel in creative texts and communications.
The new “router” will automatically determine which model is best suited for a specific question or task and will select it without additional actions from the user. However, there will still be an option to manually choose a model for those who want to control this process themselves. According to an OpenAI researcher named Rune, this feature is intended to help avoid situations where users are left with a less appropriate model for important queries.
Information about the “router” appeared in publications on X from specialists involved in AI development or analysis. It is known that the system will be able to distinguish requests that require logical thinking, creativity, or the use of external tools, and will direct each task to the appropriate model.
This change is expected to simplify the use of ChatGPT for a wide range of users who do not always know which model to choose for their request. Automatic model selection will save time and improve the quality of responses, especially in complex areas such as medicine or programming.
