Arioron/Amber-Fable-1.0

Arioron/Amber-Fable-1.0 is a 1.7 billion parameter specialized language model, fine-tuned by Arioron using LoRA on the Qwen3-1.7B base model. Engineered for mathematical reasoning and algorithmic logic, it achieves 75.0% accuracy on GSM8K and 55.0% on MATH benchmarks. This model is highly efficient for educational tools and logic-based tasks, trading some general world knowledge for peak reasoning capability.