Model fiche
Jais 30B Chat v3
By MBZUAI / Core42 · UAE
chat
general
multilingual
Overview
MBZUAI and Core42's reference open Arabic LLM — a 30B trained natively (not a fine-tune) for Arabic with strong English bilingual support under Apache 2.0.
When to pick this model
- Arabic-first chat and content generation
- Bilingual Arabic-English customer support workloads
- MENA-region sovereign deployments
- Research on non-Latin-script LLM training
- Apache-licensed Arabic alternatives to closed APIs
VRAM requirements by quantization
| Quantization | VRAM required |
|---|---|
| Q4_K_M (recommended) | 18 GB |
| Q5_K_M | 22 GB |
| Q8_0 | 33 GB |
| FP16 (no quantization) | 60 GB |
VRAM figures include model weights plus a typical 8k KV cache and ~600 MB runtime overhead (Ollama / llama.cpp baseline). Add headroom for higher context lengths.
Strengths
- Native Arabic-first architecture, not a Llama fine-tune
- Strong bilingual Arabic-English performance
- Apache 2.0 license enables commercial use
- Backed by MBZUAI, Core42, and Cerebras
Limitations
- 8k context limits long-document workflows
- No official Ollama distribution
- Weaker than Jais Adapted 70B for hardest Arabic tasks
Architecture & training
Architecture: Native decoder · SwiGLU · ALiBi · Arabic-first
Training: MBZUAI + Core42 + Cerebras.
Verdict
The default open Arabic LLM at the 30B class — pick it for native Arabic quality without the Llama license.
Quick start
# HuggingFace : core42/jais-30b-chat-v3Or use the open-source MCP server to query this model from Claude Desktop, Cursor, or any MCP-compatible client.