NbAiLab/borealis-1b

TEXT GENERATIONPricing:Input $0.04 / Cached $0.008 / Output $0.08Concurrent Unit Cost:1Model Size:1BQuant:BF16Context Size:32kPublished:May 20, 2026License:otherArchitecture:Transformer0.0K Featherless Exclusive Cold

NbAiLab/borealis-1b is a 1-billion parameter instruction-tuned model from the National Library of Norway, based on Google's Gemma 3 family. This full release model incorporates a limited subset of Norwegian rights-holder data, specifically for generating news article titles and ingress texts. It is fine-tuned for Norwegian-centric instruction following, assistant-style tasks, writing, summarization, and question answering, offering specialized performance for the Norwegian language.

Loading preview...

Borealis 1B: Norwegian-Centric Instruction-Tuned Model

NbAiLab/borealis-1b is a 1-billion parameter instruction-tuned model developed by the National Library of Norway (Nasjonalbiblioteket, NB). It is part of the Borealis family, built upon the Google Gemma 3 architecture, and specifically fine-tuned for Norwegian-centric instruction following.

Key Capabilities

  • Norwegian Language Specialization: Optimized for assistant-style tasks, writing, summarization, and question answering in Norwegian.
  • Unique Data Integration: This "full release" model incorporates a limited subset of data from an agreement between Norwegian rights-holder organizations and the Norwegian government, specifically to enhance its ability to generate news article titles and ingress texts.
  • Safety Alignment: Utilizes "prompt baking" and weighted merging of SFT and aligned models to balance quality, usefulness, and safer behavior, though users should still implement safeguards for critical applications.
  • Context Length: Supports a context length of 32768 tokens.

Good For

  • Norwegian Language Applications: Ideal for developers building applications requiring high-quality text generation and understanding in Norwegian, including Bokmål and Nynorsk.
  • Assistant-Style Tasks: Suitable for drafting, summarization, and Q&A in a Norwegian context.
  • Content Generation: Particularly useful for generating news-related content like titles and ingress texts, leveraging its unique training data.
  • Research and Evaluation: Can be used for assessing and improving Norwegian writing style and quality, and for evaluating language coverage.

Finetunes
6 models