Official documentation
Audar-ASR-V1-Turbo (model card)
AudarAI, Hugging Face, audarai organization, 2026
Read the originalRead on October 21, 2026
- What it is
- The model card AudarAI publishes on its own Hugging Face organization for Audar-ASR-V1-Turbo, its open-weight accuracy-tier speech recognition model. It states the model transcribes 30 languages in total, with Arabic dialects included.
- How much weight it carries
- Published by the maker on its own organization, so it is the maker's documentation rather than a third-party listing. It covers the Turbo tier only; the card does not give language counts for the other Audar-ASR tiers, and the audarai.com homepage, pricing and privacy pages state no count.
Facts taken from this source
- AudarAI says its Audar-ASR-V1-Turbo model transcribes 30 languages in total, covering MSA and the major Arabic dialects, code-switched Arabic and English, and English.
“It transcribes MSA and every major Arabic dialect ... across 30 languages in total.” (Overview section of the model card)