Left is faster. Higher reads more hard parts right.
ElevenLabs v4 is better on non-streaming latency, streaming latency and naturalness. MAI-Voice-2.1 is better on accuracy and price.
Last updated Methodology by Dileep Chagam
Measure by measure, the same sentences, the same place.
An estimate: speech-to-text transcribers scored it, not people.
The same sentence, read by each voice.
ElevenLabs v4. ElevenLabs v4: 678 ms. MAI-Voice-2.1: 861 ms.
ElevenLabs v4. ElevenLabs v4: 678 ms. MAI-Voice-2.1: 861 ms.
ElevenLabs v4. ElevenLabs v4: 561 ms. MAI-Voice-2.1: Can't stream.
Neither has a figure. ElevenLabs v4: Finishes first. MAI-Voice-2.1: No stop message.
MAI-Voice-2.1. ElevenLabs v4: 80.4%. MAI-Voice-2.1: 82.2%.
MAI-Voice-2.1. ElevenLabs v4: $80.00. MAI-Voice-2.1: $22.00.