anique-1/khaleeji-qwen2.5-7b-finllm

Hugging Face
TEXT GENERATIONConcurrency Cost:1Model Size:7.6BQuant:FP8Ctx Length:32kPublished:May 6, 2026License:apache-2.0Architecture:Transformer0.0K Open Weights Warm

Khaleeji-FinLLM by anique-1 is a 7.6 billion parameter, bilingual (Arabic & English) large language model built on the Qwen 2.5 7B Instruct architecture, featuring a 32K context length. It is specifically fine-tuned for the financial, legal, and economic landscape of the GCC countries, offering deep regional expertise. This model excels in financial advisory, investment insights, and regulatory compliance within the Gulf region. It is designed to serve as an agentic expert for specialized financial tasks in the Khaleej.

Loading preview...

Khaleeji-FinLLM: Gulf Region Financial Expert

Khaleeji-FinLLM, developed by anique-1, is the first multi-lingual (Arabic & English) large language model specifically fine-tuned for the financial, legal, and economic landscape of the GCC (Gulf Cooperation Council) countries. Built upon the robust Qwen 2.5 7B Instruct architecture, this 7.6 billion parameter model with a 32K context length is designed to be a premier agentic expert for the Khaleej.

Key Capabilities & Differentiators

  • First of its Kind: The initial multi-lingual model dedicated to the specific financial regulations and market dynamics of the Gulf region.
  • Bilingual Mastery: Trained on over 500,000 high-quality samples of financial data, expert Q&A, and regulatory documents in both Arabic and English.
  • Regional Expertise: Possesses a deep understanding of specific financial frameworks across Saudi Arabia (Zakat, Corporate Tax, Vision 2030), UAE (VAT, Free Zone laws, DIFC/ADGM), and Qatar & Kuwait (energy sector, banking standards).
  • Agentic Reasoning: Capable of serving as a backbone for financial AI agents in the region.

Use Cases

  • Financial Advisory: Complex analysis of Zakat and local taxation.
  • Investment Insights: Market analysis for Gulf-based energy, real estate, and fintech sectors.
  • Regulatory Compliance: Summarization and interpretation of local financial laws.

Training Details

The model was fine-tuned using LoRA (Rank 64) with subsequent full-weight merging, leveraging BFloat16 precision on AMD MI3110X (ROCm) accelerators. It is released under the Apache 2.0 license.

Finetunes
1 models