DeepSeek V4 Model Switcher

Problem Solved

In long conversations, users often want to start on DeepSeek V4 Flash for quick answers, then switch to DeepSeek V4 Pro for complex reasoning, long context, or higher-precision tasks. If each change requires manually specifying model parameters or restarting context, the workflow feels heavy. Model Switcher addresses the immediate selection of a model inside the current session: it switches between two DeepSeek V4 tiers based on natural-language intent and returns a short confirmation.

How It Works

The skill listens for switching intent, such as asking to switch Pro, switch Flash, or switch back. When triggered, it performs three steps:

• Identify whether the target is deepseek/deepseek-v4-pro or deepseek/deepseek-v4-flash
• Call session_status(model=...) to update the current session model
• Reply that the current model has changed and mention the approximate Token cost

The key point is that it only changes the current session model override. It does not modify the default model, nor does it affect other sessions. This makes model selection a lightweight conversational action rather than a configuration step.

Scope and Notes

This skill is not a general-purpose model router. It only covers the two models listed in the source: Pro is stronger but more expensive, while Flash is faster and cheaper. Switching applies only to the current session; other sessions in a group chat are not affected. Replies should stay concise, so each switch does not require verbose explanations.