Strengths supported by sources
- Supports structure and affinity workflows in one openly distributed project.jwohlwend/boltz official source · README.md: Introduction, Inference, Understanding the affinity prediction and License
model · family
Boltz
Boltz is a biomolecular interaction model family. Boltz-2 adds affinity prediction to complex-structure prediction.
At a glance
Explanatory profile: source reviewed · Automated source review, 2026-09-16. This does not change the review status of its results.
| Property | Description and evidence |
|---|---|
| Access | Repository states code and models use the MIT licencejwohlwend/boltz official source · README.md: Introduction, Inference, Understanding the affinity prediction and License |
| Model type | Not extracted or verified for this record. |
| Known versions | Not extracted or verified for this record. |
| Training data | Not extracted or verified for this record. |
| Context limits | Not extracted or verified for this record. |
| Code licence | Not extracted or verified for this record. |
| Weights licence | Not extracted or verified for this record. |
Versions and evaluated configurations
- Boltz-2 · variant of
How it works
Schematic of the documented input, computation and output; not an executable configuration.
Read the diagram as text
- Molecular inputs
- MSA / Pairformer features
- Coordinate diffusion
- Structure
- Separate affinity module
Boltz-2 architecture
The Boltz-2 implementation combines molecular and alignment features with a Pairformer module. A conditioned diffusion module predicts coordinates; a separate affinity module produces binding outputs. This architecture description applies to Boltz-2, not automatically to every Boltz family release.
jwohlwend/boltz official source · src/boltz/model/models/boltz2.py at the pinned repository revision: MSAModule, PairformerModule, DiffusionConditioning and AffinityModule; README InferenceThe documented YAML input describes the biomolecules and requested properties. Structure prediction and affinity outputs are distinct: one affinity output estimates binding strength, while another classifies binders against decoys.
jwohlwend/boltz official source · README.md: Introduction, Inference, Understanding the affinity prediction and LicenseBenchmarks and results
Release 2026-09-16-d74d282221a9 · 0 evaluations · 0 metric rows. Different protocols are not a single leaderboard.
No evaluations linked in this release.
Strengths and limitations
Limitations and conditions
- Binder probability and affinity regression are trained with different supervision and must not be compared as the same metric. Unqualified CLI calls select the latest model.jwohlwend/boltz official source · README.md: Introduction, Inference, Understanding the affinity prediction and License
Profile review details
Primary project documentation or paper inspected for the explanatory claims and cited locations. Reviewed coverage concerns this narrative, not complete metadata, independent reproduction or a performance ranking.
Stable record: discovery-model-boltzApplicable tests and references
Applicability is distinct from a completed evaluation.
- PLINDER · Proposed association
Sources and history
Release 2026-09-16-d74d282221a9 · Record review: discovered
- jwohlwend/boltz official source · Original source · b1ebfc46ecf57f5414e0d1a6f9027bbb122c53bc
Technical metadata and extraction receipts
Stable ID: discovery-model-boltz
- areas
- molecular-interactions
- access
- official_source_linked
- benchmark applicability
- candidate; not evidence of a reported evaluation
- candidate benchmark ids
- discovery-benchmark-plinder
- entity level
- family
- missing metadata
- checkpoint: unextracted; code licence: unextracted; parameters: unextracted; training cutoff: unextracted; training data: unextracted; version: unextracted; weights licence: unextracted
- reported name
- Boltz
- version
- Not reported