4 ms·

Currently, we're using GPT-4o. We've tested it with Claude as well and plan to roll out support soon!