top of page

BioMed Benchmarks

Airspeed Velocity (ASV)

ASV is a Python benchmarking tool designed to measure the performance of Python projects, including data processing libraries and frameworks.

DEBS (Distributed and Event-based Systems)

DEBS benchmarks focus on evaluating the performance of event-driven and stream processing systems.


HiBench is a benchmark suite for Hadoop-based big data platforms, covering various workloads such as MapReduce, Spark, and machine learning.

BBBP (Blood-Brain Barrier Penetration)

The BBBP dataset contains information on whether certain chemical compounds can pass the blood-brain barrier or not. It is used to evaluate AI models' capability to predict the potential of compounds to cross the blood-brain barrier, a crucial aspect in central nervous system drug development

DUD-E: The Database of Useful Decoys

DUD-E is designed to help benchmark molecular docking programs by providing challenging decoys.

InfluxDB TSM (Time Series Benchmark)

This benchmark measures the performance of time series databases like InfluxDB, focusing on time-stamped data handling.


BigBench is a benchmark suite for big data analytics systems, designed to measure the performance of SQL-based big data analytics workloads.

DeepChem Benchmarks

DeepChem is a popular open-source library for deep learning in cheminformatics. It includes several benchmark datasets and tasks for evaluating various AI models' performance in drug discovery, including molecular property prediction, virtual screening, and activity prediction


LinkBench is a benchmark designed to assess the performance of graph databases by measuring the efficiency of graph data operations.


ChEMBL is a large bioactivity database that provides bioactivity data for small molecules against different biological targets. It includes information on compound-target interactions and is commonly used to evaluate the performance of AI models in predicting compound activities


The FreeSolv dataset includes experimentally measured solvation free energies for small molecules in water. It is used to assess the accuracy of AI models in predicting solvation free energies, a critical property in drug design and pharmacokinetics


MOSES (Molecular Sets) is a benchmark dataset for molecular generative models. It includes diverse sets of molecular structures with properties like synthetic accessibility, drug-likeness, and uniqueness, enabling the assessment of AI models' ability to generate valid and novel molecules
bottom of page