aixplain

Benchmark reports you can actually explore.

Independent, interactive evaluations of LLMs and ASR models — starting with Arabic, built to extend to every language and modality we support.
Explore LLM Benchmarks →
TOP SYSTEMS · ARABIC ASR

    Arabic ASR Evaluation

    Word Error Rate

    WER, % · Lower is better

    Speed

    Latency, seconds · Lower is better

    Cost per Task

    Illustrative cents per call · Lower is better

    We provide transparent ASR evals — across every dialect we cover.

    Dialect:

    Dataset Distribution

    Dialect: