Speech Models Just Keep Improving
Canto: a speech model built for the real world | Wispr Flow
https://wisprflow.ai/canto
.jpg)
Today, the Wispr Advanced Interfaces Lab is introducing Canto, our latest speech model for real-time dictation. On an evaluation of real-world dictations, Canto achieved the lowest word error rate among all the models we tested.
Filed September 18, 2026 at 5:30 am
I switched from Wispr Flow to Gemini ( https://www.jeffturner.info/just-dug-up/a-challeger-to-wisprflow/ ) and now to the new Siri. I switched to Siri because it transcribes on the device. I didn't do it for privacy reasons. Privacy died a long time ago. I did it for consistency of experience. Connection issues sometimes plagued both Wispr Flow and Gemini.