Boltz-2 via Boltz CLI v2.2.0, 50 diffusion samples, seed 42, MSA server (Smorodina et al. 2026); Boltz CLI v2.2.0, --model boltz2

Overview

Configuration as run in the cited comparison.

Consult the linked sources for architecture or protocol details. Missing evidence is not evidence of a missing capability.

Evaluations and results

3 evaluations · 7 results. Different protocols are not a single leaderboard.

Filter evaluations

Tested model or methodProtocolDatasetEvaluation setupMetricEvidence origin

Applied filters: All linked evaluations

Exact evaluated configurations and original reported results
Tested configuration	Protocol and dataset	Finding	Evidence and details
Configuration: Boltz-2 via Boltz CLI v2.2.0, 50 diffusion samples, seed 42, MSA server (Smorodina et al. 2026)	Protocol: Telling cognate from shuffled nanobody-antigen pairs by ipTM Dataset: Smorodina et al. 2026 all-against-all VHH-antigen pairing matrix (91 unique PDB entries)	0.026 average-precision unitless · higher Uncertainty: Not reported by the source Coverage: Not reported scored / Not reported eligible	Independent external evaluation · Source checked Methods, coverage and source Boltz-2 on Telling cognate from shuffled nanobody-antigen pairs by ipTM (Smorodina et al. 2026) structural-20261009-protocol-smorodina2026-vhh-cognate-vs-shuffled Aggregation: Not reported Structural Plausibility Without Binding Specificity: Limits of AI-Based Antibody-Antigen Structure Prediction Confidence Scores · Results P11, 'AP = 0.026' for Boltz-2
Configuration: Boltz-2 via Boltz CLI v2.2.0, 50 diffusion samples, seed 42, MSA server (Smorodina et al. 2026)	Protocol: Best DockQ against sampling depth on cognate nanobody-antigen complexes Dataset: Smorodina et al. 2026 nanobody-antigen benchmark: 106 cognate VHH-antigen complexes	0.57 dockq unitless · higher Uncertainty: Not reported by the source Coverage: Not reported scored / Not reported eligible	Independent external evaluation · Source checked Methods, coverage and source Boltz-2 on Best DockQ against sampling depth on cognate nanobody-antigen complexes (Smorodina et al. 2026) structural-20261009-protocol-smorodina2026-vhh-dockq-vs-sampling Aggregation: Not reported Structural Plausibility Without Binding Specificity: Limits of AI-Based Antibody-Antigen Structure Prediction Confidence Scores · Results P36, Boltz-2 (0.57 to 0.80), N = 1
Configuration: Boltz-2 via Boltz CLI v2.2.0, 50 diffusion samples, seed 42, MSA server (Smorodina et al. 2026)	Protocol: Best DockQ against sampling depth on cognate nanobody-antigen complexes Dataset: Smorodina et al. 2026 nanobody-antigen benchmark: 106 cognate VHH-antigen complexes	0.8 dockq unitless · higher Uncertainty: Not reported by the source Coverage: Not reported scored / Not reported eligible	Independent external evaluation · Source checked Methods, coverage and source Boltz-2 on Best DockQ against sampling depth on cognate nanobody-antigen complexes (Smorodina et al. 2026) structural-20261009-protocol-smorodina2026-vhh-dockq-vs-sampling Aggregation: Not reported Structural Plausibility Without Binding Specificity: Limits of AI-Based Antibody-Antigen Structure Prediction Confidence Scores · Results P36, Boltz-2 (0.57 to 0.80), N = 100
Configuration: Boltz-2 via Boltz CLI v2.2.0, 50 diffusion samples, seed 42, MSA server (Smorodina et al. 2026)	Protocol: How well ipTM tracks DockQ on cognate nanobody-antigen complexes Dataset: Smorodina et al. 2026 nanobody-antigen benchmark: 106 cognate VHH-antigen complexes	0.665 pearson-correlation unitless · higher Uncertainty: Not reported by the source Coverage: Not reported scored / Not reported eligible	Independent external evaluation · Source checked Methods, coverage and source Boltz-2 on How well ipTM tracks DockQ on cognate nanobody-antigen complexes (Smorodina et al. 2026) structural-20261009-protocol-smorodina2026-vhh-iptm-dockq-calibration Aggregation: Not reported Structural Plausibility Without Binding Specificity: Limits of AI-Based Antibody-Antigen Structure Prediction Confidence Scores · Results P19, Boltz-2: ipTM against DockQ for the best-DockQ sample of each complex
Configuration: Boltz-2 via Boltz CLI v2.2.0, 50 diffusion samples, seed 42, MSA server (Smorodina et al. 2026)	Protocol: How well ipTM tracks DockQ on cognate nanobody-antigen complexes Dataset: Smorodina et al. 2026 nanobody-antigen benchmark: 106 cognate VHH-antigen complexes	-0.04 pearson-correlation unitless · higher Uncertainty: Not reported by the source Coverage: Not reported scored / Not reported eligible	Independent external evaluation · Source checked Methods, coverage and source Boltz-2 on How well ipTM tracks DockQ on cognate nanobody-antigen complexes (Smorodina et al. 2026) structural-20261009-protocol-smorodina2026-vhh-iptm-dockq-calibration Aggregation: Not reported Structural Plausibility Without Binding Specificity: Limits of AI-Based Antibody-Antigen Structure Prediction Confidence Scores · Results P21, Boltz-2: change in ipTM against change in DockQ under saturation sampling
Configuration: Boltz-2 via Boltz CLI v2.2.0, 50 diffusion samples, seed 42, MSA server (Smorodina et al. 2026)	Protocol: How well ipTM tracks DockQ on cognate nanobody-antigen complexes Dataset: Smorodina et al. 2026 nanobody-antigen benchmark: 106 cognate VHH-antigen complexes	18% proportion percent · lower Uncertainty: Not reported by the source Coverage: Not reported scored / Not reported eligible	Independent external evaluation · Source checked Methods, coverage and source Boltz-2 on How well ipTM tracks DockQ on cognate nanobody-antigen complexes (Smorodina et al. 2026) structural-20261009-protocol-smorodina2026-vhh-iptm-dockq-calibration Aggregation: Not reported Structural Plausibility Without Binding Specificity: Limits of AI-Based Antibody-Antigen Structure Prediction Confidence Scores · Results P19, Boltz-2: share of predictions with ipTM at least 0.5 and DockQ below 0.23 (confident failures, Q2)
Configuration: Boltz-2 via Boltz CLI v2.2.0, 50 diffusion samples, seed 42, MSA server (Smorodina et al. 2026)	Protocol: How well ipTM tracks DockQ on cognate nanobody-antigen complexes Dataset: Smorodina et al. 2026 nanobody-antigen benchmark: 106 cognate VHH-antigen complexes	0.493 pearson-correlation unitless · higher Uncertainty: Not reported by the source Coverage: Not reported scored / Not reported eligible	Independent external evaluation · Source checked Methods, coverage and source Boltz-2 on How well ipTM tracks DockQ on cognate nanobody-antigen complexes (Smorodina et al. 2026) structural-20261009-protocol-smorodina2026-vhh-iptm-dockq-calibration Aggregation: Not reported Structural Plausibility Without Binding Specificity: Limits of AI-Based Antibody-Antigen Structure Prediction Confidence Scores · Results P19, Boltz-2: ipTM against DockQ for the first sample (sample0) of each complex

Source checking is not independent reproduction. Release 2026-10-10-6e93f504adfc.

Use this model

How it works, versions and access

Evidence

Source checking verifies the cited claim or transcription. It does not establish independent reproduction.

Evidence table

Inspect claims, sources and review details

Trace each statement to its source and review. A context-only reference supports the record generally; it does not verify an individual field. Source checking does not reproduce an experiment.

One row per statement and cited source. Multiple citations are not independent evaluations. Shared locators are labelled explicitly.

0 evidence rows matching the loaded filters

Claims, original sources and review scope · Release 2026-10-10-6e93f504adfc
Property and statement	Original source and location	Review and provenance

No evidence rows match these filters. Choose another scope or clear the search.

All evidence (JSONL) (gzip)How to read the evidence

Sources and history

Release 2026-10-10-6e93f504adfc · Record review: source checked

1 source records and release history

Structural Plausibility Without Binding Specificity: Limits of AI-Based Antibody-Antigen Structure Prediction Confidence Scores · Original source · bioRxiv version 1, posted 2026-03-03; Europe PMC preprint full text PPR1221387 (manuscript EMS215481); not peer reviewed

Download this release (gzip)

Technical metadata and extraction receipts

Stable ID: structural-20261009-config-smorodina2026-boltz2

areas: proteins-complexes
contexts: research
method types: foundation_model
reported name: Boltz-2
foundation model eligible: true
source locator: Methods, structure prediction (P74)
version: Boltz CLI v2.2.0, --model boltz2
parameters: --use_msa_server; recycling_steps = 3; sampling_steps = 200; diffusion_samples = 50; seed 42
training overlap: 64 of 106 cognate systems in the training set; 42 not (Boltz-2 cutoff about 1 June 2023)

Related records