Fastest GPU For Text to Speech

Generating spoken audio, measured against realtime: 10x means ten seconds of speech every second.

Kokoro TTS 82M

Weights: hexgrad/Kokoro-82M

NVIDIA L40S243.99
NVIDIA H100 80GB HBM3222.02
NVIDIA H200170.82
NVIDIA A100 40GB SXM4142.03
NVIDIA A100 80GB SXM4110.17
NVIDIA A10G101.11
NVIDIA L497.35
NVIDIA T442.88

Measured in x realtime. Longer is faster.

#GPUx realtime
1NVIDIA L40S243.99
2NVIDIA H100 80GB HBM3222.02
3NVIDIA H200170.82
4NVIDIA A100 40GB SXM4142.03
5NVIDIA A100 80GB SXM4110.17
6NVIDIA A10G101.11
7NVIDIA L497.35
8NVIDIA T442.88

Every number is a first-party run on our own bench. Cards missing from a board have not been measured on that model yet, or cannot fit it.