📚 3LM: A Benchmark for Arabic LLMs in STEM and Code

In three linesHugging Face releases 3LM, a benchmark evaluating Arabic LLMs on STEM and code tasks. The dataset covers mathematics, physics, chemistry, and programming in Arabic, enabling measurement of model capabilities in these specialized domains.

Summary generated by Claude — human-verified