[Feature Request]: Add "Vision" capability/incapability indicators

[HumanistAtypik]

Feature Description

Add visual indicators to show which modes and which models support image analysis, and which don't.

Example:

Eye icon (👁️) for modes AND models with vision capabilities

Slashed eye icon for modes AND models without vision capabilities

Apply these indicators to each model and mode in the selection interface.

Use Case

Users can instantly identify which models can process images without failed trials and errors, nor usage-credit wasting.

Alternatives Considered

Simulate a fake "Vision" mode in a 3rd column or area on the model selection menu, listing only the few capable models: