Compare transcription models as sufficient public audio-duration data becomes available.
Weekly transcribed minutes by model once exposed.
This chart appears once public gateway aggregates expose enough data.
