# Hugging Face launches Open TTS Leaderboard for faster model evaluation

> The new leaderboard uses objective metrics to rank open‑source speech models in hours instead of weeks.

Oossa · 2026-09-30 · https://oossa.com/en/hugging-face-launches-open-tts-leaderboard-for-faster-model-evaluation

Produced and translated with AI assistance. Check the original sources below.

Hugging Face announced an Open TTS Leaderboard that rates open‑source text‑to‑speech models with automatic metrics. It measures intelligibility via word error rate, speed with real‑time factor and time‑to‑first‑audio, and speaker similarity using cosine similarity of embeddings. The system can evaluate a model in a few hours, compared with the weeks needed for human‑preference arenas. The leaderboard also lets users listen to outputs and give feedback.

## The facts

- Published Sep 30, 2026 – "PUBLISHED: Wed Sep 30 2026 02:00:00 GMT+0200"
- Ranks over 8,000 TTS models; top English WER models include hexgrad/Kokoro-82M, Supertone/supertonic-3, fishaudio/s2-pro

## Why it matters

It gives developers a quick, reproducible way to compare open‑source speech models, helping them choose faster or more accurate options for apps.

## Sources & references

1. [Open TTS Leaderboard: Scalable Evaluation for Multilingual Text-to-Speech and Voice Cloning](https://huggingface.co/blog/open-tts-leaderboard) – Hugging Face, 2026-09-30

Last updated: 2026-09-30