QCRI/AZERG-T2-Mistral

TEXT GENERATIONConcurrent Unit Cost:1Model Size:7BQuant:FP8Context Size:4kTool Calling:SupportedPublished:Jul 23, 2025License:apache-2.0Architecture:Transformer0.0K Open Weights Featherless Exclusive Cold

QCRI/AZERG-T2-Mistral is a 7 billion parameter language model developed by QCRI, fine-tuned from Mistral-7B-Instruct-v0.3. This specialist model is optimized for STIX Entity Type Identification, classifying entities within security text passages into predefined STIX types. It was trained on the QCRI/AZERG-Dataset to achieve high performance in identifying entities like Malware, Tool, or Threat-Actor. Its primary use is to extract potential STIX entities from security-related text.

Loading preview...

QCRI/AZERG-T2-Mistral: Specialized STIX Entity Identification

QCRI/AZERG-T2-Mistral is a 7 billion parameter model, fine-tuned from mistralai/Mistral-7B-Instruct-v0.3, specifically designed for Task 2: STIX Entity Type Identification. Developed by QCRI, this model excels at classifying entities found in security text into one of several predefined STIX entity types.

Key Capabilities

  • Specialized Classification: Identifies and categorizes entities (e.g., Malware, Tool, Threat-Actor) from security-related text passages.
  • Targeted Training: Fine-tuned on the QCRI/AZERG-Dataset for high accuracy in STIX entity detection.
  • Integration with AZERG Framework: Designed to work within the AZERG framework for automated threat intelligence.

Intended Use Cases

  • Threat Intelligence Analysis: Assists threat intelligence analysts by automatically assigning STIX entity types to identified entities.
  • Security Text Processing: Extracts potential STIX entities from various security documents and reports.
  • Automated Entity Extraction: Streamlines the process of identifying and categorizing critical information in cybersecurity contexts.