I've built an AI dictation tool (like many others), but this time I have taken the time to benchmark all of the 34 models that we provide, so users can actually make a qualified choice on what model should be the daily driver.
The best is that it's open source, so if you think we can improve something, please come and help out.