Categories Companies Research

Loading the SOTA2 catalog…

Explore

Categories Companies Research

Contribute

Join SOTA2 Contact

SOTA2

© 2026 SOTA2

Build in Bellevue, USA

WizWand is now SOTA2 Research

Papers Benchmarks Datasets Tasks

Text-to-Speech research benchmarks – Page 8 · SOTA2 Research

Text-to-Speech

Benchmarks

Dataset NameSOTA MethodSort

Dataset NameSOTA methodMetricTrendResultsLast Updated

AISHELL-3 (test)

4.18MOS

4

Jun 16, 2026

3.87NMOS

4

Jun 16, 2026

3.89NMOS

4

Jun 16, 2026

3.58MOS

4

Jun 11, 2026

Afrikaans AF 2h

4.21MOS

4

Jun 11, 2026

Subjective Evaluation (36 samples)

4.18NMOS

4

Jun 2, 2026

2.3WER

4

May 12, 2026

Tibetan multi-dialect speech corpus Kham dialect (test)

93.17STOI

4

Apr 21, 2026

Tibetan multi-dialect speech corpus Amdo dialect (test)

94.92STOI

4

Apr 21, 2026

Tibetan multi-dialect speech corpus U-Tsang dialect (test)

94.52STOI

4

Apr 21, 2026

Proposed Method (FS2 + Mimi)

0.003Overall RTF

4

Apr 15, 2026

0.026Error Rate

4

Apr 14, 2026

2.8Error Rate

4

Apr 14, 2026

1.9Error Rate

4

Apr 14, 2026

Balalaika corpus (test)

2.938MOS (Mean Opinion Score)

4

Apr 3, 2026

LibriTTS (unseen speakers)

4.6Naturalness MOS

4

Apr 2, 2026

68.2Similarity (SIM)

4

Mar 25, 2026

4.14rWER (%)

4

Mar 10, 2026

VoiceAssistant400K (test)

4.27UTMOS

4

Feb 26, 2026

WildSpoof TTS Track Unseen speakers 2026

3.908UTMOS

4

Feb 26, 2026

IndexTTS es (test)

0.826SS

4

Feb 26, 2026

DeepASMR-DB Intra-style ASMR → ASMR

4.29OI-MOS (ZH)

4

Feb 26, 2026

DeepASMR-DB Intra-style Normal → Normal

4.1ZH OI-MOS

4

Feb 26, 2026

30.05WER-O

4

Feb 26, 2026

14.96WER-O

4

Feb 26, 2026

Page 8 of 13

Popular datasets

LibriSpeech Seed-TTS Multilingual LEMAS-TTS IndicTTS Database

Follow for updates

Join SOTA2 Contact