Left is faster. Higher reads more hard parts right.
GPT-4o mini TTS is better on non-streaming latency and naturalness. Gemini 3.8 Flash-Lite 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.
GPT-4o mini TTS. GPT-4o mini TTS: 688 ms. Gemini 3.8 Flash-Lite: 729 ms.
GPT-4o mini TTS. GPT-4o mini TTS: 688 ms. Gemini 3.8 Flash-Lite: 729 ms.
Neither has a figure. GPT-4o mini TTS: Can't stream. Gemini 3.8 Flash-Lite: Can't stream.
Neither has a figure. GPT-4o mini TTS: No stop message. Gemini 3.8 Flash-Lite: No stop message.
Gemini 3.8 Flash-Lite. GPT-4o mini TTS: 82.7%. Gemini 3.8 Flash-Lite: 88.6%.
Gemini 3.8 Flash-Lite. GPT-4o mini TTS: $25.70. Gemini 3.8 Flash-Lite: $19.03.