CYFRAGOVPL/PLLuM-4B-chat-2512

VISIONConcurrency Cost:1Model Size:4.3BQuant:BF16Ctx Length:32kPublished:Jan 15, 2026License:apache-2.0Architecture:Transformer0.0K Open Weights Cold

The CYFRAGOVPL/PLLuM-4B-chat-2512 is a 4 billion parameter chat-optimized large language model developed by the PLLuM consortium and continued by HIVE AI, specialized in Polish with additional English data. Built upon the Gemma-3-4b-pt architecture, it is refined through extensive instruction tuning, preference learning, and advanced alignment techniques using a large-scale Polish and English text corpus. This model excels at generating contextually coherent text in Polish, offering assistance in various tasks, and is particularly strong in applications for Polish public administration.

Loading preview...

PLLuM-4B-chat-2512: Polish-Centric Chat Model

CYFRAGOVPL/PLLuM-4B-chat-2512 is a 4 billion parameter model from the PLLuM family, developed by the PLLuM consortium and HIVE AI, specifically designed for chat and general-purpose dialogue in Polish. It is based on the Gemma-3-4b-pt architecture and has undergone rigorous training with a focus on high-quality Polish and English data.

Key Capabilities

  • Polish Language Specialization: Trained on extensive, high-quality Polish text data, ensuring strong performance in Polish-language tasks.
  • Organic Instruction Tuning: Utilizes a unique, manually curated dataset of approximately 70,000 Polish "organic instructions" to enhance human-model interaction and mitigate negative linguistic transfer.
  • Preference Learning: Incorporates the first Polish-language preference corpus with around 60,000 manually annotated pairs, teaching the model correctness, balance, and safety, especially for sensitive topics.
  • Public Administration Expertise: Achieves top scores on custom benchmarks relevant to Polish public administration, making it suitable for domain-specific intelligent assistants.
  • Retrieval Augmented Generation (RAG): Designed to perform well in RAG settings, with a specific prompt format provided for document-based question answering and citation.

Good For

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

Finetunes
1 models