Genesis ML open-sources Ruby potency-prediction model and audited GenesisPotency-2026 training data
edunov · x · 2026-11-21
Genesis ML published the technical report for Ruby, its potency-prediction model, and open-sourced its carefully audited training dataset GenesisPotency-2026 under CC BY-SA 4.0, free for academic and commercial use.
- The problem: hit ID requires triaging a billion-molecule library down to a few hundred; lead optimization means picking a dozen synthesizable analogs from millions. Free-energy methods are accurate within a series but too costly for screening; ligand-based QSAR is cheap but fails to generalize to new scaffolds; docking's scoring function remains limited.
- The method: instead of generating an explicit pose and scoring it, Ruby uses the cofolding foundation model Pearl, whose trunk outputs an implicit representation of all protein–ligand token relationships — skipping costly, often inaccurate pose generation.
- The team claims state-of-the-art potency prediction across both discovery regimes, and Ruby is now in the hands of drug hunters.
Related event: Genesis Releases Open-Source Drug Potency Model Ruby(2 posts)→
More from Research
- Grady Booch claims a Riemann Hypothesis proof discovered working with Claude — Grady_Booch · 2026-11-21
- The 250,000-page Riemann proof claim: when computational verification outpaces human comprehension — SydSteyerhart · 2026-11-21
- Unverified claim: Riemann Hypothesis solved by researcher working with Claude — _vztu · 2026-11-21
- Genesis Molecular AI launches Ruby potency model, up to 1000x faster, open-sources training data — edunov · 2026-11-21
- 'All AI can do is plagiarize': model caught regurgitating a proof straight from The Book — moyix · 2026-11-21
- Experts gave AI 10% odds of cracking a Millennium Prize Problem by 2027 — two now claimed — theimposingshadow · 2026-11-21