Left is faster. Higher reads more hard parts right.
ElevenLabs v3 is better on accuracy. Deepgram Aura-2 is better on non-streaming latency, streaming latency, stopping, naturalness 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.
Deepgram Aura-2. ElevenLabs v3: 320 ms. Deepgram Aura-2: 281 ms.
Deepgram Aura-2. ElevenLabs v3: 320 ms. Deepgram Aura-2: 281 ms.
Deepgram Aura-2. ElevenLabs v3: Can't stream. Deepgram Aura-2: 126 ms.
Deepgram Aura-2. ElevenLabs v3: No stop message. Deepgram Aura-2: 40 ms.
ElevenLabs v3. ElevenLabs v3: 87.2%. Deepgram Aura-2: 77.2%.
Deepgram Aura-2. ElevenLabs v3: $40.00. Deepgram Aura-2: $30.00.