tokhey/Qwen2.5-3B-Egyptian-MCQ

TEXT GENERATIONConcurrent Unit Cost:1Model Size:3.1BQuant:BF16Context Size:32kTool Calling:SupportedPublished:Jul 10, 2026Architecture:Transformer Featherless Exclusive Cold

tokhey/Qwen2.5-3B-Egyptian-MCQ is a 3.1 billion parameter Qwen2.5-3B-Instruct model fine-tuned by tokhey, specialized in generating English multiple-choice questions. It is specifically designed to mimic the style and structure of the Egyptian Ministry of Education's General Secondary Stage examinations. This model excels at producing curriculum-aligned grammar and vocabulary questions with structured JSON outputs, including explanations and plausible distractors.

Loading preview...

Overview

tokhey/Qwen2.5-3B-Egyptian-MCQ is a specialized fine-tuned version of the Qwen2.5-3B-Instruct model, developed by tokhey. It focuses on generating English multiple-choice questions (MCQs) that adhere to the specific style and difficulty of the Egyptian Ministry of Education's General Secondary Stage examinations. The model was instruction-tuned using LoRA (PEFT) on a synthetic dataset of approximately 6,400 samples, teaching it the structure and writing style of official Egyptian English assessments rather than memorizing questions.

Key Capabilities

  • Generates Ministry-style English MCQs for grammar and vocabulary.
  • Supports topic-based and unit-based generation, aligning with curriculum content.
  • Outputs questions in a structured JSON format, including a statement, one correct answer, three plausible distractors, and an explanation for each question.
  • Optimized for production deployment, with vLLM support recommended for inference.

Intended Use Cases

This model is ideal for:

  • Educational platforms requiring authentic Egyptian Ministry-style English assessments.
  • Automatic exam generation and curriculum-aware assessment systems.
  • AI teaching assistants needing to create targeted English practice questions.

Limitations

  • Specifically optimized for Egyptian secondary education English examinations.
  • Outputs should always be reviewed by a human before official use.