QalbunLLM-V1

Qalbun AI

Qalbun AI is an innovative AI model designed to integrate Islamic principles with cutting-edge artificial intelligence. Built to respect and reflect traditional values while harnessing modern technology, Qalbun AI offers context-aware, ethical, and culturally aligned interactions.

Overview

Qalbun AI bridges the gap between timeless Islamic wisdom and state-of-the-art AI research. The model adapts to user inputs with a deep understanding of context, ensuring that each response honors the ethical and cultural dimensions of Islam.

Available Versions

Qalbun AI is offered in multiple deployment variants to meet diverse performance and resource requirements:

• Full Precision (f32):
  The most accurate version, providing the highest performance and complete model capabilities—ideal for research and critical applications.

• Half Precision (f16):
  A balanced variant that reduces memory usage while still delivering robust performance.

• Quantized 8-bit (Q8):
  An optimized model that significantly lowers memory consumption with a slight trade-off in accuracy.

• Quantized 4-bit (Q4):
  The most lightweight version, designed for resource-constrained environments while maintaining essential functionality.

Key Features

• Ethical AI:
  Developed with a commitment to Islamic ethics, Qalbun AI integrates core cultural values into every interaction.

• Contextual Understanding:
  Delivers intelligent, context-aware responses that respect both modern nuances and traditional teachings.

• Interdisciplinary Integration:
  Merges insights from AI research and Islamic studies to create a holistic, balanced approach to technology.

• Continuous Learning:
  Evolves over time by adapting to new research, user feedback, and emerging trends in both AI and ethical frameworks.

Releases

You can now run Qalbun AI offline on your phone! Download here

Author

Qalbun AI is developed and maintained by the Polyvor Labs Teams. Their multidisciplinary expertise in artificial intelligence and Islamic studies drives the innovation behind this project.

The Development Team (Polyvor Labs)

Zahir Hadi Athallah<br> Thiflul Ma'ani Minal Mu'min<br> Ruby Hardianto