Voxtral Mini 4B Realtime Arabic

Use it to transcribe Arabic speech in streaming applications such as live captions and voice interfaces.

Availability

Model checkpoint available on Hugging Face

Pricing model

Not established

Deployment

Use with vLLM or Transformers

Privacy

Not established

Licence

Apache 2.0 License

• The model supports Arabic dialects and Modern Standard Arabic transcription. ^[1]

  Voxtral Mini 4B Realtime Arabic is a streaming speech-to-text model for **Arabic dialects and Modern Standard Arabic**.

• It is designed to handle code-switching between languages within a conversation. ^[1]

  It is designed to handle challenging real-world audio and **code-switching**, where speakers alternate between languages within the same conversation.

• It processes 16 kHz audio and produces text as audio arrives. ^[1]

  It processes 16 kHz audio through a causal audio encoder and language decoder, producing text as audio arrives.

• The model can be used with vLLM or Transformers. ^[1]

  The model is co-developed with the Ministère de la Transition Numérique et de la Réforme Administrative (MTNRA) under the umbrella of the strategic partnership signed between Mistral and Morocco in January 2026 as a sovereign AI asset.

• At a configured transcription delay of 480 ms, it achieves an average CER of 8.82% across seven Arabic benchmarks. ^[1]

  At a configured transcription delay of **480 ms**, the model achieves an average Character Error Rate (CER) of **8.82%** across seven Arabic benchmarks.

• The model is licensed under the Apache 2.0 License. ^[1]

  This model is licensed under the [Apache 2.0 License](https://www.apache.org/licenses/LICENSE-2.0.txt).

• Turn a prompt into an editable slide, poster, or dashboard Weekly brief 2026-W41
• Two local models can classify, route, or score text and images Daily brief 2026-10-10