xanhnon/Qwen2.5-0.5B-Instruct-LLM

TEXT GENERATIONPricing:Input $0.04 / Cached $0.008 / Output $0.08Concurrent Unit Cost:1Model Size:0.5BQuant:BF16Context Size:32kTool Calling:SupportedPublished:Jul 31, 2026License:apache-2.0Architecture:Transformer Open Weights Featherless Exclusive Cold

The xanhnon/Qwen2.5-0.5B-Instruct-LLM is a 0.5 billion parameter instruction-tuned causal language model, fine-tuned by xanhnon on the open Qwen2.5* model. Optimized for Vietnamese language understanding and generation, it excels at tasks such as reading comprehension, information extraction, question answering, and summarization. This model demonstrates significant performance improvements over its base Qwen2.5-0.5B-Instruct counterpart on Vietnamese benchmarks like VMLU, ViSquad, ViDrop, and ViDialog, making it suitable for applications requiring strong Vietnamese NLP capabilities.

Loading preview...

Model Overview

The xanhnon/Qwen2.5-0.5B-Instruct-LLM is a 0.5 billion parameter instruction-tuned causal language model developed by xanhnon. It is part of a collection of models (including Llama3.2 and Qwen2.5 variants) that have been specifically fine-tuned for enhanced performance in Vietnamese language tasks.

Key Capabilities

  • Vietnamese Language Understanding and Generation: Optimized for native Vietnamese content.
  • Instruction Following: Designed to respond effectively to instructions for various NLP tasks.
  • Core NLP Tasks: Excels in:
    • Reading Comprehension
    • Information Extraction
    • Question Answering
    • Summarization

Performance Highlights

This model shows substantial improvements over its base Qwen2.5-0.5B-Instruct model on several Vietnamese benchmarks:

  • VMLU: Achieved 49.7, a +10.6 point increase.
  • ViSquad: Scored 87.3, a +24.8 point increase.
  • ViDrop: Reached 62.3, a +30.8 point increase.
  • ViDialog: Improved to 39.0, a +11.0 point increase.

While demonstrating strong performance in Vietnamese, the model's ranking on the ArenaHard benchmark (10.9% win rate) indicates its primary strength lies in its specialized Vietnamese optimization rather than broad multilingual or general reasoning tasks.

When to Use This Model

  • Vietnamese-centric Applications: Ideal for projects requiring high accuracy in Vietnamese text processing.
  • Resource-constrained Environments: Its 0.5B parameter size makes it efficient for deployment where computational resources are limited.

Limitations

  • May exhibit hallucinations on cultural-specific content.
  • Primary focus is on Vietnamese; performance may not be optimal for specialized technical domains or other languages.

Finetunes
1 models