raoulbia/Qwen2.5-7B-3GPP-NR

TEXT GENERATIONConcurrency Cost:1Model Size:7.6BQuant:FP8Ctx Length:32kPublished:Jan 12, 2026License:cc-by-nc-4.0Architecture:Transformer0.0K Open Weights Cold

raoulbia/Qwen2.5-7B-3GPP-NR is a 7.6 billion parameter language model, fine-tuned from Qwen2.5-7B-Instruct, specifically designed for technical questions about 3GPP 5G New Radio (NR) specifications. It excels at providing concise and accurate answers on 5G NR protocols, procedures, and configurations, based on training with nearly 27,000 Q&A pairs derived from 3GPP 38-series technical specifications. This model is optimized for assisting telecom engineers and researchers with domain-specific inquiries.

Loading preview...

Qwen2.5-7B-3GPP-NR: 5G NR Technical Expert

This model, developed by raoulbia, is a specialized fine-tuned version of the Qwen2.5-7B-Instruct base model, focusing exclusively on 3GPP 5G New Radio (NR) specifications. It was trained using LoRA on a dataset of approximately 27,000 question-answer pairs extracted from 3GPP 38-series technical documents.

Key Capabilities & Features

  • Domain Expertise: Highly proficient in answering detailed technical questions about 5G NR protocols, procedures, and configurations.
  • Concise Responses: Generates answers that are 28% shorter on average than the base model while maintaining accuracy.
  • Accurate Terminology: Uses correct 3GPP terminology (e.g., RRC, E-UTRA, ENDC, SRS, NG-RAN) and release-specific references (Rel-15, Rel-16, Rel-17).
  • Targeted Training: Fine-tuned on specific 3GPP 38-series specifications including physical layer procedures, RF requirements, conformance testing, and RRC protocol.

Intended Use Cases

  • Technical Q&A: Ideal for querying specific details within 3GPP 5G NR specifications.
  • Telecom Engineering Support: Assists engineers in quickly looking up specification details.
  • Research & Education: Useful for academic and research purposes related to 5G NR.
  • Domain-Specific Chatbots: Can serve as the core for telecommunications-focused AI assistants.

Limitations

  • Scope-Limited: Expertise is confined to 38-series (5G NR) and does not cover the entire 3GPP stack.
  • General Knowledge: May struggle with broad, definitional questions outside its specialized domain.
  • Data Origin: Training data is LLM-generated and not manually verified, which may introduce biases.

Finetunes
1 models