List your company

Urdu Mathematical Reasoning on MGSM-Urdu 250-problem (test)

3.87CorrectnessRiazi-8B
2.2062.6383.073.502Jun 24, 2026

Evaluation results

MethodLinks
Riazi-8BEvaluation Protocol=LLM-as-a-Judge, Training=CPT followed by CoT instruction tuning2026.063.873.94.144.24.194.06
Llama-8BEvaluation Protocol=LLM-as-a-Judge2026.063.513.53.853.843.863.71
Alif-8BEvaluation Protocol=LLM-as-a-Judge2026.062.272.233.522.832.672.7
Qalb-8BEvaluation Protocol=LLM-as-a-Judge2026.062.272.293.152.722.672.62