OmniGenBench: A Modular Platform for Reproducible Genomic Foundation Models Benchmarking

TL;DR

OmniGenBench is a modular, reproducible benchmarking platform that standardizes evaluation of genomic foundation models, addressing key challenges in data transparency, interoperability, and interpretability to advance genomic AI research.

Contribution

It introduces a unified, extensible platform for benchmarking GFMs, enabling consistent evaluation and fostering reproducibility in genomic AI research.

Findings

Supports over 31 open-source models

Provides automated, one-command evaluation

Addresses reproducibility and interpretability challenges

Abstract

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.