MGB Challenge

by BBC + academic consortium

BBC broadcast-media ASR + diarization challenge — multi-year evaluation series.

TL;DR

BBC broadcast-media ASR + diarization challenge — multi-year evaluation series.

Best for broadcast-media ASR with diarization; English + Arabic + Egyptian dialect. Pricing: research-only.

Category
Open source
License
—
Stars
—
Last push
—
Pricing
research-only
Platforms
Web

What it is

The MGB (Multi-Genre Broadcast) Challenge series ran 2015–2019 with BBC English broadcasts and Aljazeera Arabic. Standard broadcast-ASR + diarization evaluation. Research-only.

Best for: Broadcast-media ASR with diarization; English + Arabic + Egyptian dialect.
Watch out for: BBC license · research-only · ~1600h English BBC broadcast (MGB-1/2) · Arabic + dialectal Arabic (MGB-3/5). Cite: Bell et al., ASRU 2015.

Install / use

http://www.mgb-challenge.org/  # request form

Features

Speaker diarizationYes
Word-level timestampsNo
Streaming / real-timeNo
Languages supported4
HIPAA eligibleNo

MGB Challenge vs Whipscribe

FeatureMGB ChallengeWhipscribe
CategoryOpen sourceTranscription APIs
Pricingresearch-onlyfree beta
Speaker diarizationYesYes
Word timestampsNoYes
StreamingNoNo
Languages499
PlatformsWebWeb, API, MCP

Alternatives to MGB Challenge

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.