Compare Results
Side-by-side comparison of engine results from a multi-engine benchmark run: engine, model version, native metric, and reference-pose RMSD where computed. Scores are never combined; agreement is not evidence of binding.
Each engine keeps its native metric. GNINA docking energy, DiffDock confidence, Boltz-2 binder probability and the Chai-1 aggregate score are different quantities and are never combined into one benchmark score. Chai-1 has no affinity output at all.
A multi-engine benchmark measures agreement between models. It is not experimental validation and does not establish binding, affinity, potency, selectivity, or safety.
Loading cases…