CYFRAGOVPL/PLLuM-12B-instruct-2512

Hugging Face
TEXT GENERATIONConcurrency Cost:1Model Size:12BQuant:FP8Ctx Length:32kPublished:Jan 15, 2026License:apache-2.0Architecture:Transformer Open Weights Warm

CYFRAGOVPL/PLLuM-12B-instruct-2512 is a 12 billion parameter instruction-tuned large language model from the PLLuM family, developed by the PLLuM consortium and continued by HIVE AI. Based on Mistral-Nemo-Base-2407, it is specialized in Polish language tasks with additional English data, excelling in generating contextually coherent text and assisting with various tasks. This model is particularly optimized for applications within Polish public administration and general Polish language processing, achieving state-of-the-art results on relevant benchmarks.

Loading preview...

PLLuM-12B-instruct-2512: Polish Language Model

PLLuM-12B-instruct-2512 is a 12 billion parameter instruction-tuned model from the PLLuM family, developed by the PLLuM consortium and later by HIVE AI. It is built upon the Mistral-Nemo-Base-2407 architecture and is specifically designed for the Polish language, incorporating English data for broader generalization. The model benefits from an extensive collection of high-quality Polish and English text data, rigorous cleaning, and deduplication.

Key Capabilities

  • Specialized Polish Language Processing: Excels in generating contextually coherent text and assisting with tasks in Polish.
  • Organic Instruction Dataset: Fine-tuned on approximately 70k manually curated Polish "organic instructions" and additional programmatic and synthetic instructions, covering a wide range of human-model interactions.
  • Polish Preference Corpus: Aligned using the first Polish-language preference corpus (~60k manually annotated pairs) to ensure correctness, balance, and safety, particularly for sensitive topics.
  • Strong Performance: Achieves top scores on custom benchmarks relevant to Polish public administration and state-of-the-art results in broader Polish-language tasks.
  • RAG Optimization: Trained to perform well in Retrieval-Augmented Generation (RAG) scenarios, with a specific prompt format for document-based question answering.

Good For

  • General Language Tasks: Text generation, summarization, extraction, and question answering in Polish.
  • Domain-Specific Assistants: Particularly effective for applications in Polish public administration, legal, and bureaucratic contexts.
  • Research & Development: A robust foundation for AI applications requiring strong command of the Polish language.

Finetunes
1 models