TrialBench / HINT / TOP

Suite of benchmarks for clinical trial outcome prediction.

Kind
meta-platform
Composite
76.5
Benchmarks tracked
4
Direct-linked
0
As of
2026-05-12
Host organisation
Fu/Sun Lab, Georgia Tech + HMS
Primary contacts
Tianfan Fu, Jimeng Sun, Marinka Zitnik
Founded
2022
License model
MIT
Official site
→ site
GitHub
→ GitHub
Count methodology
Fu et al. 2022-2024: HINT (17k trials), TOP (17k), TrialBench (21k trials, 12k drugs), CT-Outcome.

Rubric

rigor
4
coverage
3
maintenance
4
adoption
3
quality
4
accessibility
5
industry_relevance
4

Breakdown

hint_trials
17,000
top_trials
17,000
trialbench_trials
21,000
task_variants
12

Notes

First rigorous ML benchmarks on trial outcomes. Limited by CTgov quality.

← Back to all initiatives

Compare:
Open comparison →