SUPERB

by Academic consortium (NTU + CMU + JHU + Meta)

Speech processing Universal PERformance Benchmark — 10 English speech tasks.

TL;DR

Speech processing Universal PERformance Benchmark — 10 English speech tasks.

Best for probing English SSL speech models across ASR, KWS, SID, ER, SD, QbE, IC, SF, PR, TTS-related tasks. Pricing: free.

Category
Open source
License
—
Stars
—
Last push
—
Pricing
free
Platforms
HuggingFace, GitHub

What it is

SUPERB defines 10 frozen-feature-extraction tasks on English speech corpora to probe self-supervised models. The standard SSL leaderboard before ML-SUPERB. Toolkit: Apache-2.0; component data licenses vary.

Best for: Probing English SSL speech models across ASR, KWS, SID, ER, SD, QbE, IC, SF, PR, TTS-related tasks.
Watch out for: Apache-2.0 toolkit · uses LibriSpeech/VoxCeleb/IEMOCAP/etc. (component licenses vary, IEMOCAP is research-only) · English-only. Cite: Yang et al., Interspeech 2021.

Install / use

git clone https://github.com/s3prl/s3prl

Features

Speaker diarizationYes
Word-level timestampsNo
Streaming / real-timeNo
Languages supported1
HIPAA eligibleNo

SUPERB vs Whipscribe

FeatureSUPERBWhipscribe
CategoryOpen sourceTranscription APIs
Pricingfreefree beta
Speaker diarizationYesYes
Word timestampsNoYes
StreamingNoNo
Languages199
PlatformsHuggingFace, GitHubWeb, API, MCP

Alternatives to SUPERB

OpenAI Whisper
OpenAI

The reference open-source multilingual ASR model from OpenAI.

OSS · MIT ★ 98.1k
whisper.cpp
Georgi Gerganov

C/C++ port of Whisper — runs on anything, from a Raspberry Pi to Apple Silicon.

OSS · MIT ★ 48.8k
faster-whisper
SYSTRAN

4× faster than reference Whisper using CTranslate2 — production sweet spot.

OSS · MIT ★ 22.3k

Whipscribe is a managed faster-whisper + whisperX service. If you want transcripts without running infrastructure, paste a URL or drop a file in the form below — you'll have a transcript in seconds.