Voxtral Darija
Open-source ASR for Moroccan Darija, handling Arabic-French-English code-switching.
About
Voxtral Darija is an automatic speech recognition model for Moroccan Darija, developed by Morocco's Ministry of Digital Transition and Management Reform with Mistral AI as part of the Digital Morocco 2030 strategy.
Built on Mistral's Voxtral architecture, it transcribes speech to text even when speakers alternate between Arabic, French and English, matching how Darija is actually spoken.
It was released open-source alongside an Arabic dialect classifier as the first milestone of the partnership, intended for public departments, startups and researchers building Darija-aware digital services.
Specs
Context window
Pricing
Free quota
API access
Weights
Released
Benchmarks
Provider-reported unless a source says otherwise. Benchmarks are marketing-adjacent by nature — treat them as directional.
Use cases
Trust score — 7/10
Scored 7/10 (metrics tier, metrics-v1). Evidence: Verified via vendor site/docs/GitHub and reputable coverage, wave 116; Trending score 70/100 from researcher assessment.