Hugging Face FluentlyLM-Prinum: A 32.5B Multilingual Causal Language Model
FluentlyLM-Prinum, developed by fluently-lm, is a 32.5 billion parameter causal language model built on the QwenForCausalLM architecture. It stands out with an extensive 131,072 token context length, enabling it to process and generate long, coherent texts. The model officially supports a diverse set of languages, including English, French, Spanish, Russian, Chinese, Japanese, and Persian.
Key Capabilities & Features
- Large Context Window: Processes up to 131,072 tokens, beneficial for complex tasks requiring extensive context.
- Multilingual Support: Officially supports 7 languages, making it versatile for global applications.
- Competitive Performance: Achieved 12th place on the Open LLM Leaderboard (as of Feb 21, 2025), with an average score of 47.22.
- GGUF Compatibility: Available in GGUF format for local deployment and use with various interfaces.
Performance Highlights
Evaluations on the Open LLM Leaderboard show strong performance across several benchmarks:
- IFEval (0-Shot): 80.90
- BBH (3-Shot): 59.48
- MATH Lvl 5 (4-Shot): 54.00
- MMLU-PRO (5-shot): 53.42
Use Cases
FluentlyLM-Prinum is suitable for a wide range of applications requiring robust language understanding and generation, especially those benefiting from a large context window and multilingual capabilities. This includes advanced chatbots, content creation, code generation (as demonstrated in the quickstart), and complex reasoning tasks.