feat: multi-model parallel comparison mode

[andrewchumchal]

Feature Request

Run the same prompt against multiple providers/models simultaneously and compare responses side-by-side.

Motivation

Users regularly want to know which model answers a question better. Today that means copy-pasting the same message into multiple tabs or apps. No desktop AI client ships this well.

Scope

• "Compare mode" toggle in the chat input
• Select 2–4 models to run in parallel
• Side-by-side response columns with model labels
• Streaming responses in all columns simultaneously
• Token count + latency shown per response
• One-click to "continue" the conversation with a chosen model

Notes

This is a high-value power user feature with low implementation complexity relative to its impact. It directly justifies switching from a single-provider app.