The voice mode there would require a new model I suppose On iOS it works but it cannot read the information from live camera for example (also likely b/c of a missing model capability)

→ View original post on X — @testingcatalog