This repository contains the results of a custom benchmark designed to evaluate the performance of open-source language models on Turkish culture questions and basic reasoning tasks.
Overview
We crafted a set of 25 questions covering:
Turkish general knowledge (e.g., capital city, national holidays, geography)
Basic arithmetic and logic puzzles
Simple calculus and string-processing tasks
