Whisper has the encoder-decoder architecture, so it's hard to run streaming efficiently, though whisper-streaming is a thing.

https://kyutai.org/next/stt is natively streaming STT.