OpenBind-0 Is Fully Open—and Points to the Data Gaps That Actually Matter

Open source means more than a runnable model

OpenBind-0 uses the Apache 2.0 license and releases training data, code, model weights and training recipes. Users can do more than run inference: they can retrain the model or experiment with proprietary data mixed into the original training distribution.^[S1][S2]

For a scientific foundation model, architecture is only part of the story. Data provenance, cutoff dates, modality mixtures, checkpoint selection and whether fine-tuning preserves the original distribution all affect behavior. Opening these components makes OB0 a baseline that can be studied and modified rather than a black-box service.^[S1][S2]

On standard protein–ligand benchmarks, OB0 is competitive with strong systems such as AlphaFold3 and Protenix. But the release is more interesting for the uneven results that follow—results closer to real discovery than another near-SOTA headline.^[S1]

A correct position is not enough; the molecule must also be chemically valid

A co-folding model may place a ligand in roughly the correct pose while producing unrealistic bond lengths, angles or local geometry. RMSD alone can count such a structure as correct, but chemically invalid geometry undermines free-energy calculations and medicinal-chemistry decisions.^[S1]

OpenBind adds inference-time chemical steering during diffusion sampling and uses PoseBusters to check ligand validity. The joint success rate for both correct pose and chemically valid geometry rises from 48% to 61%.^[S1]

It is not the flashiest architecture headline, but it exposes the gap between a pose that is approximately correct on a benchmark and a structure that downstream science can actually use.

Why did four more years of PDB data not improve everything?

OB0's training cutoff is June 2025, while OpenFold3-preview2 stops in 2021. Intuitively, nearly four additional years of PDB data should make every test case easier. In practice, the two models perform similarly across many train-set-similarity bins.^[S1]

The team then examined each test complex's distance to its nearest training structure. PDB added many protein–ligand structures, yet for many test cases the nearest neighbor did not become meaningfully closer. The dataset grew, but those targets did not become more familiar to the model.^[S1]

Clear gains appear mainly when new data actually change a target's position in the training distribution—when a previously unfamiliar structure becomes more in-distribution. Data value therefore cannot be measured only by structure count; the key question is which chemical and structural space it fills.^[S1]

Real projects expose three different failure modes

OpenBind evaluates models on real fragment-to-hit campaigns. The results are uneven, and that unevenness is the most useful part. On EV-A71 2A protease, OB0 reaches 92.2% top-25 success and 73.8% top-1. Related fragment structures were already in PDB before OB0's cutoff, giving the model highly relevant structural information.^[S1]

On FatA, OB0 falls to 28.2% top-25 success; the best model, Protenix-v1-20250630, reaches only 39.4%. Models usually identify the correct pocket and a reasonable protein conformation but often miss the ligand pose. The failure is not where the ligand should go, but how it should sit there.^[S1][S4]

On RdRp, models often miss the pocket itself

On Dengue and Zika RdRp, nearly every model fails. The best top-25 success rates are only 7.7% and 6.7%. RdRp is large, has multiple surface pockets and is highly flexible around the primary site; both systems are also very dissimilar to known structures in the 2021 and 2025 training sets.^[S1][S5]

Models frequently choose the wrong pocket; even when they find it, receptor conformation or ligand pose can still be wrong. Protein–ligand prediction is therefore not one uniform task: EV-A71 has relatively good coverage, FatA is dominated by ligand placement, and RdRp exposes pocket selection, conformational modeling and OOD generalization at once.^[S1]

The same class of co-folding model falling from above 90% to single digits reveals more about where it fails than an aggregate benchmark score.

Can fine-tuning close the gap? It depends on the target

The team fine-tuned models on target-specific fragment structures. On EV-A71, related data had clearly improved OpenFold3-preview2; OB0 then surpassed the fine-tuned model after absorbing those fragments into pre-training. In this system, incorporating relevant data into pre-training may be more effective.^[S1]

The result does not generalize. FatA improves only modestly, and both RdRp systems remain poor after fine-tuning. The careful conclusion is not that pre-training always beats fine-tuning, but that training tricks may not substitute for missing experimental coverage.^[S1]