Every date here is measured, not asserted. Each entry’s repository is checked against the GitHub API for its last push.Checking repository activity…
Open Drug Discovery & Development Consortium · Benchmark Index
Every open benchmark in drug discovery, indexed and independently checked.
The evaluation landscape for AI-driven drug discovery is scattered across dozens of
repositories, consortium sites and frozen challenge pages. This is an index of all of
it — what each benchmark measures, who maintains it, whether anyone still is, and,
for every single entry, the known weakness its own documentation will not tell you.
Open to anyone working on evaluation in drug discovery — academic, industry or independent. No fee, no exclusivity.
On each entry: updated < 3 months < 1 year > 1 year repo not found no repository (challenge / web resource)
Join the O3DC consortium
The Open Drug Discovery and Development Consortium is open to anyone working on how drug discovery is evaluated. Tell us who you are and we'll be in touch.
Please enter your first name.
Please enter your last name.
Please tell us where you work or study.
Independent researchers are welcome; put "Independent".
Please enter a valid email address.
Your institutional or company address, where possible.
Please confirm before submitting.
We use these details only to contact you about O3DC. They are not sold, and not shared outside the consortium's organisers. To be removed, email mmaigym@insilicomedicine.com.
✓
Thanks, we've got it
Member sign-in
Already a member? Enter the email you joined with and we'll send you a one-time sign-in link. No password needed.
Propose a benchmark, dataset-with-defined-evaluation, or leaderboard for the Index. It enters a review queue; an organiser checks it before it goes live. We index benchmarks, not models, and every entry carries a known caveat.
Give the benchmark a name.
Choose a category.
Add the lead author or maintainer.
Use an owner/name path, or leave blank.
Every entry needs a citable source URL.
Describe what it measures (80+ characters).
Add at least one task label.
The caveat is required (60+ chars) and cannot say "no known limitations".
Every entry carries one. "No known limitations" will fail the automated checks.