Skip to main content
model-family passport · Review date not recorded

BoltzGen

All-atom generative binder design across proteins, peptides and molecular partners.

2/7Evidence fields documented
60-SECOND EVALUATION VIEW

What should a scientist know before using BoltzGen?

SupportedEvidence supports the stated context with explicit boundaries
Best suited forGeneration · Optimization
Evidence supportsBinder-design evaluations: Developer / experimental reports
Evidence does not establishUniversal superiority, therapeutic success, clinical utility or regulatory acceptance.
Major limitationPerformance depends on the evaluation dataset and operating conditions.
Current registry recordBoltzGen1 recorded release · Review date not recorded. A newer version is not assumed to be universally better.

What it is

BoltzGen extends the Boltz ecosystem from complex prediction and affinity toward target-conditioned all-atom binder generation.

Evidence trail

BioAtlas keeps the path from source to decision visible. A connection records provenance; it does not imply that evidence is sufficient for every context.

Sources2 connectedPrimary resources and normalized claims
Claims1 normalizedDe novo protein or binder design
EntityBoltzGenmodel-family · BoltzGen
ReviewReview date not recordedReview date not claimed
ConclusionContext requiredAdd to an evaluation before operational use

Model passport

Entity typemodel-family
OrganizationMIT / Boltz team
Model family introduced2025
AccessOpen source
Commercial useAllowed / verify checkpoint terms
DeploymentSelf-hosted
ComputeGPU recommended
Domainsdesign
Biology → representation → computation → evidence

How BoltzGen represents biology

model-familydesign

Category is navigation. These fields describe the model-specific computational transformation and deliberately override broad category defaults.

1 · Biological inputs
Target structure or complex contextBinder-design objective
2 · Input representation
All-atom coordinatesSequence and molecular tokens
3 · Internal representation
Geometric latent representation
4 · Architecture
Generative all-atom model
5 · Learning objective
Conditional generative modelling
6 · Output representation
All-atom coordinatesSequences

Biological scale

moleculeprotein

Modalities & tasks

ProteinPeptideSmall moleculeGenerationOptimization

Registry, claims and frontier intelligence

Versioned registry

BoltzGen

1 version record · latest curated year 2025. Model-family identity remains separate from capability and access changes.

Explore version lineage →

Inputs and outputs

Inputs

Target structure or complex contextBinder-design objective

Outputs

Candidate bindersCandidate complex structures

Scientific and technical profile

Scientific principles

All-atom generative designTarget-conditioned generation

Technology

Generative all-atom modellingBoltz ecosystem
Ideas before algorithms

Scientific lineage

Explore all foundations

These are transparent concept matches—not claims that one scientist alone caused this model. Each connection is based on the model’s recorded domain, scientific principles, technical terms or an explicit lineage link.

Biologics & genome engineering

Phage display and selection of binding proteins

George P. Smith & Sir Gregory P. Winter

Display-based selection created an experimental search engine for protein binders and remains a core validation partner for computational antibody design.

Matched concepts: binder, affinity
Physical chemistry

Intermolecular forces and excluded volume

Johannes D. van der Waals

Modern force fields, docking scores, molecular dynamics and ligand–protein packing depend on these non-covalent interactions.

Matched concepts: all-atom
Structural biology

First atomic structures of globular proteins

John Kendrew & Max Perutz

Protein structure prediction and structure-based design became meaningful because experimental crystallography established the target reality to predict against.

Matched concepts: complex

Evaluation evidence

Dataset or evaluationBinder-design evaluations
Task or metricDesign success and structural quality
Evidence statusDeveloper / experimental reports
Open source ↗

Task-specific evidence only; not comparable as a universal leaderboard score.

De novo protein or binder design

Binder-design evaluations

BoltzGen · Split details not yet normalized
experimental

A structured benchmark claim is recorded; consult the linked source for numeric values and protocol details.

Claim caveats
  • Protocol, split and implementation details must match before comparing this claim with another result.

Known limitations

  • Performance depends on the evaluation dataset and operating conditions.
  • Task-specific benchmark results should not be compared across unlike domains.
  • Outputs require task-specific scientific and experimental validation.

Milestones

2025

Adds a design branch to the Boltz lineage.