McGill-NLP/AfriqueQwen-8B

Model Overview

McGill-NLP/AfriqueQwen-8B is an 8 billion parameter causal language model developed by McGill-NLP, forming part of the AfriqueLLM suite. It is built upon the Qwen3-8B-Base architecture and has undergone extensive continued pre-training (CPT) on approximately 26 billion tokens of multilingual data. This adaptation significantly improves its performance on 20 African languages while preserving its capabilities in high-resource languages like English, French, Portuguese, and Arabic.

Key Capabilities

• Multilingual Proficiency: Adapted for 20 African languages (e.g., Swahili, Hausa, Yoruba, Zulu) and maintains strong performance in 4 high-resource languages.
• Robust Base Model: Leverages the Qwen 3 8B architecture, noted for its strong performance and long-context task capabilities.
• Extensive Training Data: Continued pre-training on a diverse corpus including African monolingual data (22.8B tokens), code (1B tokens), mathematics (~1B tokens), and synthetic data.
• Long Context Window: Supports a native context length of 32,768 tokens.
• Performance Improvement: Demonstrates a significant performance increase of +25.6 (76.5%) on an overall multilingual benchmark compared to its base model, Qwen3-8B.

Good for

• Applications requiring African language support: Ideal for tasks involving text generation, understanding, or translation in any of the 20 supported African languages.
• Multilingual NLP research: Provides a strong foundation for further research and development in low-resource language processing.
• Developers seeking Qwen 3-based models with enhanced African language capabilities: Offers a specialized alternative to the base Qwen 3 models for specific regional needs.