Research labs and institutions driving Arabic NLP forward.
Organization
Focus
Key Contributions
Link
AUB MIND Lab
Foundational Arabic NLP models
AraBERT, AraGPT2, AraELECTRA
CAMeL Lab, NYUAD
Arabic NLP tools and models
CAMeLBERT, camel_tools
UBC-NLP
Dialectal Arabic, multimodal models
MARBERT, AraT5, NileChat, PEARL, Dallah
QCRI
Arabic LLMs, text processing
Fanar LLMs, AraDiCE, Farasa
SILMA AI
State-of-the-art Arabic LLMs
SILMA LLMs, Arabic Broad Benchmark
MBZUAI
Multimodal and speech models
AIN, ArTST, ClArTTS
ARBML
Democratizing Arabic NLP
masader, klaam, tkseem
NAMAA-Space
Arabic NLP models & dialect hub
Qari-OCR, EgypTalk-ASR, Masrawy translator, GLiNER Arabic
ITIDA / MCIT (Egypt)
Egypt's national AI, sovereign models
Karnak LLM, BelMasry, Torgoman, SIA, AcQua
Omartificial-Intelligence-Space
Arabic embedding models
GATE, Matryoshka embeddings
TII (Technology Innovation Institute)
Arabic LLM benchmarks, Falcon
Open Arabic LLM Leaderboard, Falcon LLM
SDAIA (Saudi Data & AI Authority)
Sovereign Arabic LLM, national AI strategy
ALLaM model, SADA dataset, BALSAM benchmark
HUMAIN (Saudi PIF)
Full-stack AI company
ALLaM 34B, HUMAIN Chat, AI factories
KAUST
AI research, Arabic NLP
Center of Excellence in Generative AI
KFUPM-JRCAI
Joint SDAIA-KFUPM AI research
Arabic AI text detection datasets
Prince Sultan University (RIOTU)
Arabic language models
ArabianGPT, Arabic IoT AI
SinaLab, Birzeit University
Arabic NLP tools and datasets
SinaTools, Wojood NER
G42 / Inception AI
Arabic-centric LLMs
Jais LLM family
FreedomIntelligence
Arabic LLMs and alignment
AceGPT, Arabic cultural datasets
Helsinki-NLP
Machine translation models
OPUS-MT Arabic translation models
LightOn AI
Arabic web data
ArabicWeb24 corpus (39B+ tokens)
π Benchmarks & Leaderboards
Standardized evaluation frameworks for Arabic language models.
Benchmark
Description
Link
MTEB Arabic Leaderboard
Massive Text Embedding Benchmark for Arabic
Arabic Broad Leaderboard (ABL)
NextGen evaluation for Arabic LLMs by SILMA AI
Open Arabic LLM Leaderboard
Evaluation of Arabic LLMs across multiple benchmarks
ALUE
Arabic Language Understanding Evaluation benchmark
BALSAM
Benchmark of Arabic Language AI Systems and Models
SILMA RAGQA Benchmark
Evaluates Arabic/English LMs in Extractive QA tasks
Arabic Broad Benchmark (ABB)
Comprehensive evaluation tool for Arabic LLMs
ArabicMMLU
Multi-task language understanding from school exams
GATmath and GATLc
Benchmarks from Saudi GAT exams
ArabicRAGB
Arabic RAG Benchmark (multi-dialect)
ACVA
Arabic Cultural Value Alignment (8000+ questions, 58 areas)
AraDiCE
Benchmarks for dialectal and cultural capabilities of LLMs
Open Universal Arabic ASR Leaderboard
Multi-dialectal Arabic speech recognition benchmark
KITAB-Bench
Arabic OCR benchmark: 8,809 samples, 9 domains, 36 sub-domains (MBZUAI)
π€ State-of-the-Art Models
π¬ Large Language Models (LLMs)
Arabic-centric and multilingual LLMs with strong Arabic capabilities.
Model
Params
Developer
Key Features
Link
Jais
13B, 30B
Inception AI, Cerebras
Arabic-centric, bilingual, instruction-tuned
SILMA 1.0
9B
SILMA AI
Top-ranked Arabic LLM built on Gemma
ALLaM
7B
SDAIA & IBM
Saudi's sovereign model, enterprise-focused
Fanar-1-9B
9B
QCRI
Arabic-English LLM
AceGPT
7B
FreedomIntelligence
Top performance, culturally aligned
Atlas-Chat
2B-27B
MBZUAI-Paris Lab
Moroccan Darija dialect
NileChat-3B
3B
UBC-NLP
Egyptian and Moroccan dialects
Nile-Chat
4B-12B
MBZUAI-Paris Lab
Egyptian Arabic and Arabizi scripts
AraGPT2
1.5B
AUB MIND Lab
GPT-2 for Arabic text generation
Command R7B Arabic
7B
Cohere
Arabic-optimized Command R variant
SambaLingo-Arabic
7B, 70B
SambaNova
Arabic-adapted Llama 2
ArabianGPT
0.1B
Prince Sultan University
GPT-2 for Arabic, RIOTU Lab
Nuha
-
Elm (Saudi)
Multi-modal Arabic-first LLM for gov services, dialect-aware
Llama 3.3
70B
Meta
Strong Arabic performance
Qwen 3
0.6B-235B
Alibaba
Multilingual with Arabic support
Gemma 3
1B-27B
Google
Multimodal capabilities
Cohere Command-A
111B
Cohere
Optimized for RAG
Mistral Saba
24B
Mistral
Commercial API
Falcon Arabic
7B
TII (UAE)
First Arabic model in Falcon series, top of OALL
Falcon-H1-Arabic
3B-34B
TII (UAE)
Hybrid Mamba-Transformer, 128K-256K context
ALLaM-2
7B-70B
SDAIA & IBM
500B+ Arabic tokens, largest Arabic training set
Mulhem
-
SDAIA
Open-source Arabic-first LLM from Saudi Arabia
Arcee-Meraj
72B
Arcee AI
Enterprise Arabic LLM based on Qwen2-72B
Arcee-Meraj-Mini
7B
Arcee AI
Top OALL among 7B models
Aya-Expanse
8B-32B
Cohere
State-of-the-art multilingual with strong Arabic
SILMA Kashif
2B
SILMA AI
Lightweight RAG-optimized Arabic model
Ar-stablelm-2-chat
1.6B
Stability AI
Small Arabic chat model
Yehia
7B
Navid-AI
Based on ALLaM, Arabic instruction-tuned
AraLLaMA
7B
Bashar Alhafni
LLaMA2 pre-trained on Arabic data
Shahin
14B
malhajar
Syrian Arabic dialect, Qwen2.5-based
Labess Chat
7B
Linagora
Tunisian Arabic, based on Jais architecture
Phi-4
4B
Microsoft
Multilingual with Arabic, compact & efficient
ALLaM 34B
34B
HUMAIN (Saudi)
Most advanced Arabic LLM, 8PB training data, powers HUMAIN Chat
Karnak
30B-70B
ITIDA (Egypt)
Egypt's national sovereign LLM, top Arabic in its class
NOOR
10B
TII (UAE)
World's largest Arabic NLP model at launch, GPT-3 architecture
Noon
7B
Naseej
BLOOM-based Arabic LLM, instruction-tuned
Fanar Star
7B
QCRI (Qatar)
Trained from scratch on 1T Arabic/English tokens
JASMINE
0.3B-6.7B
UBC-NLP
Arabic GPT for few-shot learning, 400GB training data
Kawn
-
Misraj AI (Saudi)
Arabic-first ecosystem with 15-dialect support
Kuwain
1.5B
Misraj AI
Arabic SLM via language injection, 70% cost reduction
Mutarjim
1.5B
Misraj AI
Arabic-English translation, rivals GPT-4o mini
GemmAr
7B
ClusterlabAi
Gemma Arabic instruction-tuned on InstAr-500k
LlamAr
8B
ClusterlabAi
LLaMA 3 Arabic instruction-tuned on InstAr-500k
Calme 2.2
72B
MaziyarPanahi
Qwen2-based, strong OALL performance
Model
Params
Developer
Key Features
Link
AIN
8B
MBZUAI
Arabic-centric Large Multimodal Model
Dallah
-
UBC-NLP
Advanced multimodal LLM for Arabic
Peacock
7B
UBC-NLP
Arabic multimodal (InstructBLIP + AraLLaMA)
Qalam
-
-
Arabic OCR/HWR multimodal LLM, 0.80% WER
π§ Transformer-based Models
Model
Developer
Key Features
Link
AraBERT
AUB MIND Lab
First BERT for Arabic, multiple versions
AraBERTv02
AUB MIND Lab
Improved tokenization (135M params, 12M+ downloads)
CAMeLBERT
CAMeL Lab, NYUAD
MSA, Dialectal, and Classical Arabic
MARBERT
UBC-NLP
Focused on Dialectal Arabic and MSA
MARBERTv2
UBC-NLP
Updated with improved dialectal coverage
AraELECTRA
AUB MIND Lab
ELECTRA for Arabic
AraT5
UBC-NLP
T5 for Arabic summarization, translation, paraphrasing
SaudiBERT
King Saud University
Saudi dialect BERT, trained on 141M tweets (STMC) + forums
EgyBERT
faisalq
Egyptian dialect BERT, trained on 34M tweets + 44M forum sentences
SA-BERT-V1
Omartificial-Intelligence-Space
Saudi Arabic BERT, fine-tuned MARBERTv2
Model
Developer
Key Features
Link
GATE-AraBert-v1
Omartificial-Intelligence-Space
SOTA on MTEB Arabic STS
Arabic-Triplet-Matryoshka-V2
Omartificial-Intelligence-Space
Matryoshka representation
Swan
UBC-NLP
Dialect-aware, cross-lingual
asafaya/bert-base-arabic
asafaya
BERT-based Arabic embeddings
ModernBERT-Arabic
BounharAbdelaziz
ModernBERT-based sentence embeddings
DIMI-embedding
AhmedZaky1
Matryoshka + AraBERT for NLI
π― Task-Specific Models
Model
Task
Key Features
Link
CAMeLBERT-MSA-Sentiment
Sentiment Analysis
Fine-tuned for MSA sentiment
t5-arabic-summarization
Summarization
T5 for Arabic news summarization
opus-mt-en-ar
Translation ENβAR
Helsinki-NLP (3.5M+ downloads)
opus-mt-ar-en
Translation ARβEN
Helsinki-NLP (12.4M+ downloads)
arabic-gec-v1
Grammar Correction
Gemma-3-1b for Arabic GEC
Arabic-Text-Correction
Text Correction
AraT5-based text correction
arat5-dialects-translation
DialectβMSA
AraT5 dialect translation
Masrawy Translator
ENβEgyptian Arabic
150K+ rows, 10M+ tokens, Egyptian dialect
GLiNER Arabic
Named Entity Recognition
Plug-and-play Arabic NER (NAMAA-Space)
π£οΈ Speech Recognition (ASR)
Model
Key Features
Link
openai/whisper-large-v3
Supports Arabic among many languages
MasriSwitch-Gemma3n
Egyptian Arabic code-switching transcription
wav2vec2-large-xlsr-53-arabic
Fine-tuned on Common Voice & Arabic Speech Corpus
artst_asr_v3
ArTST for ASR on MGB2 (best for MSA)
EgypTalk-ASR-v2
High-performance ASR for Egyptian Arabic
whisper-large-v3-turbo
4x faster Whisper, Arabic support, MIT license
NVIDIA FastConformer Arabic
SOTA Arabic ASR, 115M params, 760h training, CC-BY-4.0
NVIDIA FastConformer Arabic (Diacritics)
Arabic ASR with diacritical marks, 1100h training
SeamlessM4T v2
Meta's all-in-one ASR + translation, ~100 languages inc. Arabic
MMS-1b-all
Meta's Massively Multilingual Speech, ASR for 1100+ languages
SpeechBrain wav2vec2 Arabic
Wav2Vec2 + CTC fine-tuned on CommonVoice 14 Arabic
HuBERT-Large Arabic
HuBERT fine-tuned on 2000h Arabic, 17.68% WER
HuBERT Egyptian Arabic
HuBERT fine-tuned for Egyptian dialect on MGB-3
Whisper Quran
Whisper fine-tuned for Quranic recitation recognition
Whisper Egyptian Arabic
Whisper-medium fine-tuned on 72h Egyptian speech
Whisper Arabic (small)
Whisper-small fine-tuned for Arabic language
KalemaTech Arabic STT
Whisper-small optimized for Arabic STT
Voxtral Mini
Mistral's speech model, 3B, Arabic support, Apache 2.0
Klaam
Arabic ASR/TTS/classification library (MSA + Egyptian)
π Text-to-Speech (TTS)
Model
Key Features
Link
facebook/mms-tts-ara
Facebook's Massively Multilingual Speech
speecht5_tts_clartts_ar
SpeechT5 for Classical Arabic TTS
F5-TTS-Arabic
F5-TTS with regional diversity
Arabic-F5-TTS-v2
Improved F5-TTS Arabic fine-tune
Arabic-TTS-Spark
Spark TTS fine-tuned on 300h clean Arabic, MSA with diacritics
LLMVoX
30M streaming TTS for any LLM, first Arabic autoregressive TTS
XTTS-v2
Coqui's multilingual voice cloning, Arabic support, 6s cloning
Arabic MMS Speech Synthesis
VITS-based Arabic TTS from MMS (36M params)
tts-arabic-pytorch
Tacotron2 + FastPitch + HiFi-GAN for Arabic
tts_arabic (ONNX)
FastPitch + Mixer-TTS in ONNX for offline Arabic TTS
Arabic Tacotron TTS
End-to-end Arabic TTS based on Tacotron
Piper TTS
Fast local neural TTS, Arabic voices available
π Arabic TTS Datasets & Benchmarks
Resource
Description
Link
ArVoice
Multi-speaker MSA corpus, 83h, 11 voices, diacritized
ClArTTS
Classical Arabic TTS corpus by MBZUAI
Arabic Speech Corpus
Nawar Halabi's Levantine Arabic TTS corpus (CC-BY-4.0)
SawtArabi
First Arabic dialectal + code-switching TTS benchmark
Open-Source Arabic TTS Benchmark
SILMA AI's auditory assessment benchmark
π Optical Character Recognition (OCR)
Model
Key Features
Link
QARI-OCR v0.2
Updated Qwen2-VL 2B for Arabic OCR; WER 0.160, CER 0.061
Qari-OCR-0.1-VL-2B-Instruct
Qwen2 VL 2B fine-tuned for Arabic OCR
Baseer
Arabic document-to-markdown, Qwen2.5-VL-3B based (Misraj AI), WER 0.25
Arabic-English-handwritten-OCR-v3
Qwen2.5-VL-3B, trained on 47K handwriting samples
AtlasOCR
First Darija/Moroccan Arabic OCR, Qwen2.5-VL-3B based
DIMI-Arabic-OCR
Printed Arabic with diacritics OCR (Qwen2-VL based)
arabic-large-nougat
End-to-end structured OCR for Arabic documents
arabic-small-nougat
Smaller Nougat variant for Arabic document OCR
π§ Traditional OCR Tools (Arabic Support)
Tool
Key Features
Link
PaddleOCR
High-performance multilingual OCR with Arabic support
EasyOCR
Ready-to-use OCR with Arabic support (80+ languages)
Tesseract OCR
Open-source OCR engine with Arabic language packs
π Arabic OCR Datasets & Benchmarks
Resource
Description
Link
KITAB-Bench
8,809 samples across 9 domains & 36 sub-domains (MBZUAI, ACL 2025)
SARD
Synthetic Arabic OCR dataset for recognition training
Misraj-DocOCR Benchmark
Arabic document OCR evaluation benchmark by Misraj AI
Model
Key Features
Link
blip-Arabic-flickr-8k
BLIP fine-tuned for Arabic image captioning
βοΈ Diacritization (Tashkeel)
Models and tools for adding harakat (ΨΨ±ΩΨ§Ψͺ) to Arabic text.
Model / System
Key Features
Link
CATT
Character-based Tashkeel Transformer, SOTA results
Fine-Tashkeel
Fine-tuned ByT5, 40% WER reduction
Sadeed
Small language model for diacritization
Shakkala
Neural vocalization using bidirectional LSTM
Mishkal
Rule-based diacritizer with dictionary lookups
Dataset
Description
Link
Tashkeela
Arabic diacritization corpus
arabic-text-diacritization
Benchmark dataset with systems comparison
π£οΈ Dialect Identification
Resources for identifying and classifying Arabic dialects.
Task
Description
Link
NADI 2025
Multidialectal Arabic Speech Processing (8-way dialect + ASR)
NADI 2024
Fifth Nuanced Arabic Dialect Identification
NADI Shared Tasks
Ongoing series of Arabic DID shared tasks
Dataset
Description
Link
QADI
Twitter-based multi-class dialect classification
Arabic POS Dialect
POS tagging in Arabic dialects
Arabic Dialects to MSA
Parallel dialect-MSA corpus
Casablanca
Multidialectal Arabic speech dataset (NADI 2025)
Dataset
Description
Link
masader
Largest public catalogue of Arabic NLP datasets (600+)
101 Billion Arabic Words
Massive Arabic web corpus
ArabicWeb24
39B+ tokens of high-quality Arabic web content
ArabicText-Large
743K articles for LLM training
Arabic Billion Words
Abu El-Khair corpus: 5M+ articles, 1.5B+ words
Arabic Tweets
41GB+ of Arabic tweets (~4B words)
Wojood
Nested NER corpus (550K tokens)
CIDAR
Culturally relevant instruction dataset (10K pairs)
Arabic_Function_Calling
First Arabic function calling dataset (50K+ samples)
ArabicaQA
Large-scale Arabic Question Answering
Mixed Arabic Datasets (MAD)
Community-driven diverse Arabic texts
Arabic-OpenHermes-2.5
Arabic OpenHermes instruction dataset
Alpaca Arabic Instruct
Arabic Alpaca instruction dataset
Rasaif
Classical Arabic-English parallel texts (24 books)
Shifaa Medical
Arabic medical consultation dataset
Shifaa Mental Health
Arabic mental health consultations
Arabic Reasoning Dataset
9.2K instruction-based reasoning QA pairs
Gazelle
Arabic writing assistance dataset
arabic-hate-speech-superset
Comprehensive hate speech detection
ArabicCorpus2B
1.9B word Arabic corpus
The Arabic E-Book Corpus
1,745 books (81.5M words)
BAREC Corpus
Arabic Readability Assessment
palm
Human-created Arabic instruction dataset
dialogue-arabic-dialects
Levantine, Egyptian, Gulf dialect dialogues
STMC
Saudi Tweets Mega Corpus (141M+ Saudi dialect tweets)
KFUPM Arabic AI Text Detection
Machine-generated Arabic text across multiple LLMs
Egyptian Dialogue
4,322 Egyptian Arabic-English parallel pairs from TV subtitles
NileChat LHV-Egypt
Egyptian dialect dataset for LLM pre-training
NileChat Arabizi-Egypt
Egyptian Arabizi dataset for LLM pre-training
Dataset
Description
Link
ClArTTS
Classical Arabic TTS dataset
Arabic Speech Corpus
South Levantine Arabic (Damascian accent)
Arabic-English Code-Switching
Code-switching speech from YouTube
Egyptian Arabic ASR Clean
~72 hours of Egyptian Arabic speech
MADIS5
Spoken Arabic dialects
SADA
Saudi Audio Dataset: 668h from 57 TV shows, multi-dialect (SDAIA)
SADA22 (MSA)
MSA subset of SADA, Khaliji speech
SawtArabi
Arabic speech dataset
πΌοΈ Vision & Multimodal Datasets
Dataset
Description
Link
PEARL
Multimodal Culturally-Aware Arabic Instruction Dataset
Arabic-Image-Captioning_100M
100 million Arabic image captions
Calliar
Online Arabic calligraphy (2500 samples)
arabic-img2md
15K PDF pages paired with Markdown
Arabic-OCR-Dataset
1M+ Arabic OCR samples
Arabic-VLM-Full-Pearl
309K multimodal examples for VLM training
π§ Essential Tools & Libraries
βοΈ Toolkits & Preprocessing
Tool
Description
Link
camel_tools
Suite of Arabic NLP tools (morphology, POS, NER, etc.)
Farasa
Fast and accurate Arabic text processing toolkit
SinaTools
Open source toolkit by SinaLab (Python APIs, CLI)
Qalsadi
Arabic morphological analyzer and lemmatizer
PyArabic
Python package for Arabic text manipulation
tnkeeh
Arabic text cleaning, normalization, preprocessing
Maha
Text processing library for Arabic text
Mishkal
Arabic text diacritizer (rule-based)
arabicprocess
Python library for Arabic preprocessing
MADAMIRA
Morphological analysis, diacritization, POS tagging
π Specialized Libraries
Library
Task
Link
klaam
Speech Recognition, Classification, TTS
tkseem
Arabic Tokenization
arabic-stop-words
Largest list of Arabic stop words
qawafi
Arabic poetry analysis
arabic_vocalizer
Deep-learning diacritization (ONNX format)
Tool
Description
Link
opus-mt-en-ar
English β Arabic neural MT
opus-mt-ar-en
Arabic β English neural MT (12.4M+ downloads)
π Research Papers & Conferences
π€ Conferences & Workshops
Conference
Year
Link
ArabicNLP 2025
2025
ArabicNLP 2024
2024
ArabicNLP 2023
2023
Arabic NLP Winter School
2025
AbjadNLP Workshop
2025-2026
OSACT
Ongoing
π Foundational & Survey Papers
The Landscape of Arabic Large Language Models (2025) [1]
AraBERT: Transformer-based Model for Arabic Language Understanding (Antoun et al., 2020) [2]
The Interplay of Variant, Size, and Task Type in Arabic Pre-trained Language Models (Inoue et al., 2021) [3]
ARBERT & MARBERT: Deep Bidirectional Transformers for Arabic (Abdul-Mageed et al., 2021) [4]
Jais and Jais-chat: Arabic-Centric Foundation and Instruction-Tuned LLMs (Sengupta et al., 2023) [5]
Wojood: Nested Arabic Named Entity Corpus and Recognition (Jarrar et al., 2022) [6]
Deep Learning for Arabic NLP: A Survey (Al-Ayyoub et al., 2018) [7]
Evaluating Arabic LLMs: A Survey of Benchmarks, Methods, and Gaps (Alzubaidi et al., 2025) [8]
π Recent Papers (2024-2025)
Swan and ArabicMTEB β Dialect-Aware, Cross-Lingual Language Understanding (Bhatia et al., 2024) [9]
GATE β General Arabic Text Embedding for Enhanced STS (Nacar et al., 2025) [10]
A Survey of LLMs for Arabic Language and its Dialects (Mashaabi et al., 2024) [11]
Hate speech detection in Arabic β corpus design and evaluation (2024) [12]
NADI 2024 β Fifth Nuanced Arabic Dialect Identification Shared Task (2024) [13]
CATT β Character-based Arabic Tashkeel Transformer (2024) [14]
ArabianGPT β Native Arabic GPT-based LLM (2024) [15]
SambaLingo β Teaching LLMs New Languages (2024) [16]
SaudiBERT β BERT Pretrained on Saudi Dialect Corpora (2024) [17]
ALLaM β Large Language Models for Arabic and English (SDAIA & IBM, 2024) [18]
EgyBERT β BERT Pretrained on Egyptian Dialect Corpora (2024) [19]
π Tutorials & Learning Resources
Resource
Description
Link
Arabic NLP Winter School
Two-day intensive at MBZUAI (Jan 2025)
ArabicNLP Conference
Annual ACL-affiliated conference by SIGARAB
AbjadNLP Workshop
NLP for languages using Arabic script
SDAIA Academy
Saudi AI & data science training programs
KAUST Academy (AI Specialization)
Free AI training for Saudi nationals
Resource
Description
Link
Hugging Face NLP Course
Free NLP course (applicable to Arabic models)
AUB MIND Lab Arabic-NLP Demo
Interactive demo for Arabic NLP tasks
MoroccoAI Darija Resources
Curated Moroccan Arabic dialect NLP resources
NNLP-IL Arabic Resources
Comprehensive Arabic NLP resource list
ποΈ Research Institutions
Academic and government research labs driving Arabic NLP innovation.
Institution
Country
Focus
Link
KAUST (CEMSE)
πΈπ¦ Saudi Arabia
Generative AI center, Arabic NLP research, sentiment analysis
SDAIA-KFUPM Joint Research Center (JRCAI)
πΈπ¦ Saudi Arabia
Arabic AI text detection, Arabic NLP datasets
Prince Sultan University (RIOTU Lab)
πΈπ¦ Saudi Arabia
ArabianGPT, Arabic IoT/robotics AI
King Saud University
πΈπ¦ Saudi Arabia
SaudiBERT, Saudi dialect corpora (STMC, SFC)
Nile University
πͺπ¬ Egypt
AI research, M.Sc. in AI co-designed with MIT/IBM
Ain Shams University
πͺπ¬ Egypt
Arabic NLP, sentiment analysis, NER research
MBZUAI
π¦πͺ UAE
AIN, Atlas-Chat, ArabicMMLU, KITAB-Bench
CAMeL Lab (NYU Abu Dhabi)
π¦πͺ UAE
CAMeLBERT, camel_tools, morphological analysis
UBC-NLP
π¨π¦ Canada
MARBERT, AraT5, JASMINE, Swan, Palm dataset
AUB MIND Lab
π±π§ Lebanon
AraBERT, AraGPT2, AraELECTRA
QCRI
πΆπ¦ Qatar
Farasa, Fanar LLM, Arabic POS tagging
SinaLab
π΅πΈ Palestine
SinaTools, Wojood NER corpus
ARBML
π Community
masader (600+ datasets), klaam, tkseem, qawafi
π Companies & Startups
Companies and organizations building Arabic AI/NLP products and services.
π¦πͺ United Arab Emirates
Company
Focus
Notable Products
Link
G42
AI holding company, Arabic LLMs
Jais LLM, enterprise AI solutions
Inception AI
Arabic-centric foundation models
Jais model family (with Cerebras)
Technology Innovation Institute (TII)
Open-source LLMs, research
Falcon LLM family
Saal.ai
Cognitive AI solutions
Arabic NLP, speech, generative AI
Arabot
Conversational AI for Arabic
Arabic NLP chatbot engine
Arabic.AI (Tarjama)
Arabic-first autonomous AI
Pronoia Arabic LLM, Agentic AI platform
Company
Focus
Notable Products
Link
SDAIA
Sovereign AI, national data authority
ALLaM model, SADA dataset, NCAI, BALSAM benchmark
HUMAIN
PIF-backed full-stack AI company
ALLaM 34B, HUMAIN Chat, 8PB Arabic training data
Elm
Digital transformation, gov AI (PIF-backed)
Nuha Arabic LLM, legal AI assistant, gov platform
Misraj AI
Arabic-first AI ecosystem
Kawn LLM, Baseer OCR, Mutarjim, Workforces, SeamlessAPI
Mozn
Enterprise AI, Arabic NLU
OSOS Arabic NLU platform, FOCAL compliance suite
Unifonic
Conversational AI platform
Arabic-first CX Intelligence, AI chatbots
Lucidya
AI customer experience analytics
Arabic social listening, sentiment analysis
Hudhud AI
Arabic conversational AI (no-code SaaS)
Saudi-accent chatbots, Arabic-first customer engagement
Hazen.ai
AI traffic safety & computer vision
Deep learning road safety, seatbelt/phone detection
Future Look ITC (FLITC)
Arabic-native AI solutions, venture studio
LABEAH, Smart Hire, Rayee Media, Nabadat
NAMAA-Space
Arabic NLP models hub
Qari-OCR, EgypTalk-ASR, Masrawy translator, GLiNER Arabic NER
Wittify.ai
Conversational AI for Arabic
Interactive Arabic AI agents
π‘ Saudi Arabia's AI ecosystem : Saudi Arabia leads Arabic LLM development with 53+ Arabic language models identified by Q1 2025. Key initiatives include HUMAIN ($5B+ AWS partnership, NVIDIA AI factories), Vision 2030 AI Strategy (targeting 20,000 AI specialists, 300+ AI startups, $20B investment), and SDAIA Academy/SAMAI (training 1M citizens in AI).
Company
Focus
Notable Products
Link
ITIDA / MCIT
National AI authority, sovereign models
Karnak LLM, SIA AI tutor, AcQua NLP, BelMasry, Torgoman
Intella
Arabic speech AI intelligence
Arabic STT across 25+ dialects (95.7% accuracy), Ziila digital human
WideBot AI
Arabic-first conversational AI
AQL Arabic LLM, chatbots, voicebots, AI agents
DXwand
Generative AI for Arabic business
ORXTRA platform, Arabic dialect chatbots, 20+ LLM support
Crowd Analyzer
Arabic social media monitoring
Arabic NLP analytics, sentiment analysis, media monitoring
Synapse Analytics
AI for financial inclusion
ML-powered credit scoring, Arabic data analytics
Convertedin
AI marketing automation
Arabic/English e-commerce personalization, $3M funded
Kngine
Semantic search & NLP
Arabic semantic search, data mining, knowledge engine
Monta AI
Enterprise AI solutions
LLM & RAG-based Arabic business automation
Elves
Conversational commerce
Arabic AI-assisted concierge, human-in-the-loop ML
π‘ Egypt's AI ecosystem : Egypt unveiled Karnak , its national LLM (30-80B params, highest-ranking Arabic LLM in its class), and launched the National AI Strategy 2025-2030 targeting $42.7B in AI economic value, 30,000 AI specialists, and 250+ AI startups. Karnak powers SIA (AI education tutor), AcQua (NLP call auditing), BelMasry (colloquial Arabic NLP engines), and Torgoman (specialized translation).
π International (with Arabic focus)
Company
Country
Focus
Notable Products
Link
SILMA AI
-
Arabic-first LLMs
SILMA LLMs, Arabic Broad Benchmark
Cohere
Canada
Multilingual LLMs
Command R Arabic, RAG optimization
Mistral AI
France
Multilingual LLMs
Mistral Saba (Arabic-optimized)
SambaNova Systems
USA
Arabic language adaptation
SambaLingo Arabic models
Mawdoo3
Jordan / Saudi
Arabic AI & content, NLP toolkit
Arabic LLMs, largest Arabic website, Saudi expansion
Maqsam
Jordan / MENA
Arabic speech AI, call center AI
Arabic dialect STT, AI voice bots, surpasses Google/Microsoft
Contributions are welcome! Please open an issue or submit a pull request to add new resources or update existing ones.
Fork the repository
Add your resource in the appropriate section
Ensure links are valid and descriptions are concise
Submit a pull request
This work is licensed under a Creative Commons Attribution 4.0 International License .
[1] The Landscape of Arabic Large Language Models. (2025). arXiv preprint arXiv:2506.01340 . https://arxiv.org/html/2506.01340v1
[2] Antoun, W., Baly, F., & Hajj, H. (2020). AraBERT: Transformer-based Model for Arabic Language Understanding. OSACT . https://aclanthology.org/2020.osact-1.2/
[3] Inoue, G., et al. (2021). The Interplay of Variant, Size, and Task Type in Arabic Pre-trained Language Models. WANLP . https://aclanthology.org/2021.wanlp-1.10/
[4] Abdul-Mageed, M., et al. (2021). ARBERT & MARBERT: Deep Bidirectional Transformers for Arabic. ACL . https://aclanthology.org/2021.acl-long.551/
[5] Sengupta, N., et al. (2023). Jais and Jais-chat. arXiv:2308.16149 . https://arxiv.org/abs/2308.16149
[6] Jarrar, M., et al. (2022). Wojood: Nested Arabic Named Entity Corpus. arXiv:2205.09651 . https://arxiv.org/abs/2205.09651
[7] Al-Ayyoub, M., et al. (2018). Deep learning for Arabic NLP: A survey. Journal of Computational Science , 26. https://www.sciencedirect.com/science/article/pii/S1877750317303757
[8] Alzubaidi, A., et al. (2025). Evaluating Arabic LLMs: Benchmarks, Methods, and Gaps. arXiv:2510.13430 . https://arxiv.org/abs/2510.13430
[9] Bhatia, G., et al. (2024). Swan and ArabicMTEB. arXiv:2411.01192 . https://arxiv.org/abs/2411.01192
[10] Nacar, O., et al. (2025). GATE: General Arabic Text Embedding. arXiv:2505.24581 . https://arxiv.org/abs/2505.24581
[11] Mashaabi, M., et al. (2024). A Survey of LLMs for Arabic. arXiv:2410.20238 . https://arxiv.org/abs/2410.20238
[12] Hate speech detection in Arabic. (2024). Frontiers in AI . https://www.frontiersin.org/journals/artificial-intelligence/articles/10.3389/frai.2024.1345445/full
[13] NADI 2024: Fifth Nuanced Arabic Dialect Identification. (2024). ArabicNLP . https://arxiv.org/abs/2407.04910
[14] CATT: Character-based Arabic Tashkeel Transformer. (2024). arXiv:2407.03236 . https://arxiv.org/abs/2407.03236
[15] ArabianGPT: Native Arabic GPT-based LLM. (2024). arXiv:2402.15313 . https://arxiv.org/abs/2402.15313
[16] SambaLingo: Teaching LLMs New Languages. (2024). arXiv:2404.05829 . https://arxiv.org/abs/2404.05829
[17] SaudiBERT: A Large Language Model Pretrained on Saudi Dialect Corpora. (2024). arXiv:2405.06239 . https://arxiv.org/abs/2405.06239
[18] ALLaM: Large Language Models for Arabic and English. (2024). arXiv:2407.15390 . https://arxiv.org/abs/2407.15390
[19] EgyBERT: A Large Language Model Pretrained on Egyptian Dialect Corpora. (2024). arXiv:2408.03524 . https://arxiv.org/abs/2408.03524
If you find this resource helpful, please give it a β
Made with β€οΈ for the Arabic NLP community