Short answer
AskTuring evaluates your prompt and selects the most suitable model for text and image tasks. You can override this choice at any time.
Full explanation
Automatic model routing is a feature that lets you work without worrying about which model to choose. The system examines your prompt to understand whether it contains text, images, or both. It then directs each part of the prompt to the most appropriate model.
If your prompt is text only, AskTuring chooses a language model based on the type of request, expected response time, and your preferences. If your prompt contains an image or asks for an image, AskTuring routes it to GPT Image 1, Midjourney, or NanaBanana. If your prompt contains both text and image components, AskTuring can split the request so each part goes to the right model and then combine the results into one response.
You can override model selection for a specific session or for an individual prompt. The system maintains context and memory even when switching between models. It also predicts latency to avoid sending your request to a model that is slow or unavailable. The routing framework is designed so AskTuring can add more providers in the future.
Tips
Use model overrides only when you need a specific output style or provider.
Allow automatic routing for faster responses and better performance.
Provide feedback when prompted so AskTuring can improve model selection for your workflow.
Examples
A marketing firm asks for a social media caption. AskTuring detects a text only request and routes it to a fast language model. The team can still force a specific model if they want a different writing style.
A hospital network sends a combined prompt that includes a patient education document and a diagram. AskTuring splits the prompt so text analysis goes to a language model and the diagram processing goes to an image model.
An engineering firm works in a session where they always prefer a specific model. They set a session level override, and AskTuring follows that preference while still maintaining context across responses.