Remix.run Logo
solenoid0937 an hour ago

It's probably just that the voice models are the same underlying model being served with a different system prompt and with thinking turned off.