McGill-NLP/AfriqueQwen3.5-9B-50Langs

VISIONConcurrent Unit Cost:1Model Size:9BQuant:FP8Context Size:32kTool Calling:SupportedPublished:Aug 11, 2026License:cc-by-4.0Architecture:Transformer Open Weights Featherless Exclusive Cold

McGill-NLP/AfriqueQwen3.5-9B-50Langs is a 9 billion parameter causal language model from the AfriqueLLM suite, based on Qwen 3.5. It is specifically adapted for 50 African languages through continued pre-training on approximately 35.5 billion tokens, while maintaining strong performance in high-resource languages. This model excels in multilingual contexts, particularly for African languages, and supports a native context length of 262,144 tokens. It is optimized for a broad range of language tasks across its supported languages.

Loading preview...

AfriqueQwen3.5-9B-50Langs: Multilingual LLM for African Languages

McGill-NLP's AfriqueQwen3.5-9B-50Langs is a 9 billion parameter causal language model, part of the AfriqueLLM suite, built upon the Qwen 3.5 architecture. This model is uniquely distinguished by its extensive adaptation to 50 African languages through continued pre-training (CPT) on approximately 35.5 billion tokens of multilingual data, including African monolingual, code, and mathematics datasets.

Key Capabilities

  • Broad Multilingual Support: Adapted for 50 African languages, alongside strong performance in high-resource languages like English, French, Portuguese, and Arabic.
  • Enhanced Performance: Achieves significant improvements over its base Qwen 3.5-9B model on various African language benchmarks, including AfriMGSM, AfriMMLU, and FLORES (eng->xxx), with an overall score of 62.02, representing a +15.3% increase.
  • Extended Context Window: Features a native context length of 262,144 tokens, enabling processing of very long inputs.
  • Robust Training: Utilizes a carefully curated training corpus and UniMax sampling to balance language distribution, capping high-resource languages and upsampling lower-resource ones.

Good For

  • Applications requiring strong language understanding and generation in a wide array of African languages.
  • Research and development focused on improving LLM performance in low-resource language settings.
  • Tasks benefiting from a large context window, such as document analysis or long-form content generation in supported languages.

Finetunes
1 models