🇵🇭 FilBench - Can LLMs Understand and Generate Filipino?

In three linesFilBench is a benchmark to evaluate LLM understanding and generation of Filipino. The dataset covers classification, QA, and generation tasks in Filipino and English. Results show significant gaps in major models' Filipino capabilities.

Summary generated by Claude — human-verified