Full threadravetcofx·I'd be more curious to see the performance over local models like LLaVa etc.View on HN