ElevenLabs introduces Scribe v3 and Scribe v3 Live speech-to-text models

ElevenLabs introduced Scribe v3 and Scribe v3 Live, which it calls its most accurate speech-to-text models yet. It says they are built for real-world audio and handle:

• extreme background noise;
• telling pauses from the ends of sentences;
• tracking who said what in a busy meeting.

These are ElevenLabs's claims; the post gives no benchmark figures.