OpenDDE: open-source all-atom drug discovery engine hits 51% on hard antibody-antigen benchmark
teortaxesTex · x · 2026-11-11
Aureka AI Research released OpenDDE, an open-source all-atom biomolecular foundation model positioned as a shared structural reasoning layer for AI-driven drug discovery. On the harder PXM-22to25 antibody-antigen benchmark it scores 51.0% success (DockQ ≥ 0.23), second only to NovaAtom-Lite's 66.05% and well above ESMFold2 (42.9%), Protenix-v1 (40.2%) and Boltz-1 (19.6%). The paper claims IsoDDE-level co-folding accuracy, identifies two scaling-law directions for co-folding models, and releases training code, inference pipelines, checkpoints and benchmarks on GitHub and Hugging Face — making frontier biomolecular intelligence fully reproducible and open.
More from Embodied
- CoRL 2026's wildest moments: 'Why do you care about delay?' and 'haptic doesn't matter' — breadli428 · 2026-11-11
- JKU Linz's new Institute for Machine Intelligence hires 2 PhD students in robot learning, deadline in a week — HildeKuehne · 2026-11-11
- Darpinian: Tesla should wait for the AI5 chip before scaling Cybercab production — Darpinian · 2026-11-11
- Maestro paper: cheap Gemini 2.5 beats π₀.5 baselines at zero-shot robot control — dineshjayaraman · 2026-11-11
- Tesla demos Optimus walking inside a world model at CoRL, real policy in control — Scobleizer · 2026-11-11
- NUS Proposes WM-PPO: World-Model-Predicted Successor States Improve PPO Policy Learning — NationalUniversityofSingapore · 2026-11-11