OmniGenBench: A Modular Platform for Reproducible Genomic Foundation Models Benchmarking
This provides a foundational infrastructure for reproducible genomic AI research, addressing critical reproducibility challenges in the field.
The authors tackled the lack of rigorous and reproducible evaluation for Genomic Foundation Models (GFMs) by developing OmniGenBench, a modular benchmarking platform that enables standardized, one-command evaluation across five benchmark suites and integrates over 31 open-source models.
The code of nature, embedded in DNA and RNA genomes since the origin of life, holds immense potential to impact both humans and ecosystems through genome modeling. Genomic Foundation Models (GFMs) have emerged as a transformative approach to decoding the genome. As GFMs scale up and reshape the landscape of AI-driven genomics, the field faces an urgent need for rigorous and reproducible evaluation. We present OmniGenBench, a modular benchmarking platform designed to unify the data, model, benchmarking, and interpretability layers across GFMs. OmniGenBench enables standardized, one-command evaluation of any GFM across five benchmark suites, with seamless integration of over 31 open-source models. Through automated pipelines and community-extensible features, the platform addresses critical reproducibility challenges, including data transparency, model interoperability, benchmark fragmentation, and black-box interpretability. OmniGenBench aims to serve as foundational infrastructure for reproducible genomic AI research, accelerating trustworthy discovery and collaborative innovation in the era of genome-scale modeling.