Showing Requested vs Actual Model in Chat Responses

[mohd-riaz]

Showing Requested vs Actual Model in Chat Responses

I opened a PR that adds a small transparency improvement to Odysseus.

When the model that actually responds differs from the model the user selected, the UI shows both:

• selected-model -> actual-model
• openrouter/free -> resolved-model
• primary-model -> fallback-model

If the requested and actual models are the same, the UI continues showing only a single model label.

This makes fallback chains, provider-side routing, and automatic failover much easier to understand at a glance without checking logs or debugging information.

PR is open for review and feedback. (#2562)

Screenshots