AI for Bio · Model

AlphaProteo

Closed de novo protein binder design system; seven targets reported successful and an eighth failed

ProteinsBig-tech labUnited KingdomMaintained only中文

Overview

NameAlphaProteo · aliases: AlphaProteo
Full nameSources use AlphaProteo only; no expanded name
Ambiguous names (not auto-linked)AlphaFold
FamilyNot a family entry. One technical-report version only; no family members
OrganizationGoogle DeepMind, Francis Crick Institute
Organization typeBig-tech lab
CountryUnited Kingdom
Model classDesign (de novo binder)
DomainProteins (also: Molecule generation)
First release† partly verified
Basis: Developer blog post (the technical report on arXiv, 2024-09-12, is listed separately)
Note: The blog date comes from the existing note and search results; the page was not fetched this round source
Other milestones
  • First public evidence: † · Developer blog post (the technical report on arXiv, 2024-09-12, is listed separately) source
  • Preprint v1: † · Technical report arXiv v1 (date carried from the existing note) source
  • Official announcement: † · Developer blog post (the technical report on arXiv, 2024-09-12, is listed separately) source
Latest update† partly verified · no later event
Basis: Only the release event is sourced; no new version, weights or successor found in 2025-2026
No later sourced event found (as of 2026-11-16); whether the arXiv v1 report (2024-09-12) counts as an update is a policy question and is unverified.
Note: Must never be phrased as 'no updates since' source
The most recent sourced event, not a claim that nothing changed since.
StatusMaintained only · Closed; only the 2024-09 release event is sourced
OpennessCode: Closed
Weights: Closed
Closed: no code and no weights; still no release plan as of 2026-04 (second-hand)
Benchmarks reported0 counted · none of the 44
Wet-lab validation3 recorded campaigns
Unverified fields† latest_update, first_release, wetlab_validation[1], training_data.b_private_licensed, training_data.c_own_wetlab_for_training
As of2026-11-16

What it does

Given a target structure and optional hotspot residues it generates and filters candidate binder sequences, aimed at one design round plus a small wet-lab test. It was among the first big-lab design models to publish wet-lab data, and Latent-X/X2, BoltzGen and BindCraft all use it as a comparator, but it is closed and cannot be reproduced externally.

Benchmarks reported

None of the 44 benchmarks. This model is on the zero-pairing list.

How it evaluates itself: Seven-target binder wet-lab campaign; Success-rate comparison; Follow-on functional tests

Other evaluations (outside the 44)

  • Seven-target binder wet-lab campaign (Wet-lab validation) Self-reported: One round of medium-throughput screening with self-reported better affinity and success than prior methods; VEGF-A succeeded and TNFa failed
  • Success-rate comparison (Comparison against baselines only) Self-reported: Per-target success compared with the best prior methods; the BHRF1 value is internally inconsistent and unverified
  • Follow-on functional tests (Case study) partner: The release blog describes VEGF-A and SARS-CoV-2 RBD binders tested in partner-lab assays

Developers report: Developers report that designs against seven targets, including the difficult VEGF-A, bound with 3- to 300-fold better affinity and higher success than the best prior methods after a single screening round, without iterative optimisation (self-reported).

Independent evidence: No independent evidence found. Closed and not externally runnable; later competitors only cite its numbers as comparison

Training data

Group in Figure 7: Public data + wet lab for validation only.

ClassValueDetail
a. Public databasesYesDistilled PDB and predicted-structure data; only the source databases are named, with no size, cutoff or filtering rules source
b. Private or licensed dataNot disclosedThe report does not mention private or licensed data source
c. Own wet-lab data used for trainingNot disclosedWhether own experimental data entered training is not disclosed source
d. Wet lab used only for validationYesYes: the developer's wet lab ran binding tests on eight targets, with partner functional assays source

Wet-lab validation

Every rate carries its numerator and denominator. Claims that could not be verified are shown as unverified, without numbers; different paper versions are separate entries.

Eight-target binder campaign (self-reported) (blog plus technical report)
Targets 8 · Designs tested — · Hits 7 · Success binders on 7 of 8 targets (TNFa is the failed eighth target) · Best affinity developers report 3- to 300-fold better affinities than prior methods (self-reported)

Denominator basis: Eight targets attempted (BHRF1, SARS-CoV-2 RBD, IL-7Ra, PD-L1, TrkA, IL-17A, VEGF-A and TNFa) in one round of medium-throughput screening
Self-reported with no independent replication; TNFa must be described as the failed eighth target and never merged into the seven successes

Self-reported · † · verified · source
BHRF1 per-target binding rate (unverified) (blog)
Targets 1† · Designs tested — · Hits — · Success — · Best affinity —

Denominator basis: The body text and a figure caption of the same developer source give different per-target rates, so no figures are given (unverified)
The body text is the authoritative reading, but the source is internally inconsistent, so this record marks it partial and quotes no value

Self-reported · † · †partly verified · source
Partner-lab functional assays (blog)
Targets — · Designs tested — · Hits — · Success — · Best affinity —

Denominator basis: Qualitative, no counts: some RBD binders blocked infection in cells
Partner-lab executed; not an independent replication

partner-lab · † · verified · source

Background

  • Organization: Google DeepMind is Alphabet's AI research unit (two teams merged in 2023), based in London and funded internally rather than externally; its protein design team previously built the structure-prediction series. A UK research institute contributed the wet-lab work.
  • Funding: Internally funded by Alphabet; no external funding rounds
  • Products and partnerships: Not commercialised; the developer says it is also exploring applications with a sister drug-design company.
  • Related models: AlphaFold 3, AlphaMissense, Latent-X / Latent-X2, BindCraft

Reading notes

  1. The eight targets were self-chosen and the figures are within-target hit rates, not comparable to open-competition rates.
  2. The TNFa failure shows it is no universal solution; the seven successes must be read together with that eighth failure.
  3. Being closed means no third party can re-test it, so every figure stays self-reported.
  4. Only source databases are disclosed, with no scale or cutoff, so overlap with the evaluation targets cannot be checked.

Caveats

  1. Zero pairings: it reports no score on any site benchmark.
  2. No entry or submission record in the Adaptyv competition / BenchBB; the model is closed and cannot be run externally.
  3. The BHRF1 per-target rate is internally inconsistent within one source and is treated as unverified.
  4. Training data is disclosed only at database level and must not be written up as 'public data only'.

Sources

  1. AlphaProteo technical report (arXiv 2409.08022) (preprint) · †
  2. Developer release blog (company blog) · †

Corrections

3 corrections
FieldWasNowBasisFound
wetlab_validationTNFa was presented as one of the seven successful targetsTNFa is the failed eighth target and is listed apart from the seven successessource2026-11-16
wetlab_validationThe BHRF1 per-target rate was published as a single valueBody text and figure caption disagree, so it is marked partial and no figure is givensource2026-11-16
latest_updateThe latest update was phrased as there being no updates since releaseChanged to 'no later sourced event found (as of 2026-11-16)' with event kind none-since-releasesource2026-11-16

Part of the AI for Bio benchmarks survey · all 112 models · open in the explorer. Record as of 2026-11-16.