click any column header to sort

TTS Bench — Speed — 2026-06-05_1115

Rig: windows-5090 — AMD Ryzen 9 9950X3D 16-Core Processor (16C) · NVIDIA GeForce RTX 5090 32GB · 126 GB RAM · Windows 11

Label: default voice · dramabox

TTFA = time to first audio (ms; lower is better). RTFx = realtime speed (× realtime; higher is better; e.g. 10× means 10 sec of audio generated per 1 sec of compute). Cold = first run after process start; warm = subsequent runs.

Speed winners

Fastest predefined-voice: no data

Fastest cloning-capable: DramaBox (cuda) — 2.58× warm RTFx, 3.78s warm TTFA

Model	Device	TTFA cold	TTFA warm	RTFx cold	RTFx warm	Peak RAM	Peak VRAM	Size	Released
DramaBox	cuda	5.14s	3.78s	1.93×	2.58×	2.36 GB	17.39 GB	3.3B	Apr 2026