QCRI/AZERG-T1-Mistral

TEXT GENERATIONConcurrent Unit Cost:1Model Size:7BQuant:FP8Context Size:4kTool Calling:SupportedPublished:Jul 23, 2025License:apache-2.0Architecture:Transformer Open Weights Featherless Exclusive Cold

QCRI/AZERG-T1-Mistral is a 7 billion parameter language model, fine-tuned from Mistral-7B-Instruct-v0.3, specifically designed for STIX Entity Detection. Developed by QCRI, this model excels at identifying CTI-related entities such as malware names, threat actors, and vulnerabilities within unstructured security text. It was trained on the QCRI/AZERG-Dataset to provide high performance in specialized cybersecurity intelligence tasks, making it a specialist model for the AZERG framework.

Loading preview...

Model Overview

QCRI/AZERG-T1-Mistral is a specialized 7 billion parameter language model, fine-tuned from mistralai/Mistral-7B-Instruct-v0.3. Its primary purpose is STIX Entity Detection within cybersecurity threat intelligence.

Key Capabilities

  • Specialized Entity Extraction: Designed to identify specific CTI-related entities, including:
    • Malware names
    • Threat actors
    • Vulnerabilities
  • Targeted Training: Fine-tuned on the QCRI/AZERG-Dataset to optimize performance for this specific task.
  • Integration with AZERG Framework: Developed as a specialist component for high performance within the AZERG framework.

Intended Use Cases

This model is ideal for developers and analysts who need to:

  • Automate STIX entity extraction from security text passages.
  • Enhance threat intelligence analysis by programmatically identifying key CTI elements.
  • Integrate specialized entity detection into cybersecurity platforms or workflows.

Citation

If you utilize this model in your research or applications, please cite the associated paper:

@article{lekssays2025azerg,
  title={From Text to Actionable Intelligence: Automating STIX Entity and Relationship Extraction},
  author={Lekssays, Ahmed and Sencar, Husrev Taha and Yu, Ting},
  journal={arXiv preprint arXiv:2507.16576},
  year={2025}
}