Jais Speech

Released
2026-10-15
Last refreshed
2026-10-16
Status
Researched today
Open sourceCommercial use: permittedMultimodalVision

Jais Speech is available now for vision with open-source; evaluate it while provider pricing coverage matures.

Core42 releases · 7 in the last 12 monthsChangelog →
Specifications
Family
Jais
Released
2026-10-15
Parameters
5.7B
Specialization
speech-recognition
Openness
Open source
License
Apache 2.0OSI-approvedCommercial use: permitted
Weights
Available
Code
Unknown
Created by
Core42

Sovereign AI initiative in the UAE

Mohammed Bin Zayed City, Abu Dhabi, UAE
Founded 2023
Website
Pricing

No tracked provider pricing is available yet.

Links
Website HuggingFace

About

Jais Speech (Hugging Face: IFM/Jais-ASR) is the Arabic speech-recognition model in IFM and Inception42's Jais 3 release, October 15, 2026. It transcribes Modern Standard Arabic and dialects from 19 Arab countries, keeps English or French code-switched words in their original script, identifies the dialect, and can add Arabic vowel marks. Apache 2.0 open weights. It is an audio-in, text-out model, not a general text-generation model.

Providers

No active provider routes are tracked for this model yet.

Capabilities

MultimodalAudio

Benchmark peer barsfor Vision

No task-mapped benchmark peers are available for this model yet.

Benchmark scores(1)

Scores are benchmark-specific and are direction-aware: the same numeric gap can mean very different outcomes across suites. Use the leaderboard context and this model's provider route to decide whether the winning margin is meaningful for your workload.
BenchmarkScoreVersionEvaluationSource
Open Universal Arabic ASR Leaderboard (average WER)22.9macro-average WER, six test sets (leaderboard eval script, self-run)Observed 2026-10-15—Source

Migration checks

No linked migration route is available for this model yet.

API versions

IFM/Jais-ASR