AI for Bio · Model

TxGemma

A family of drug-discovery LLMs (three sizes, predict and chat variants) instruction-tuned from an open general base on a public therapeutics dataset library

Small moleculesBig-tech labUnited Kingdom, United StatesMaintained only中文

Overview

NameTxGemma · aliases: TxGemma, txgemma-9b-predict
Full nameThe name is a product name; the model card gives no expanded form
Ambiguous names (not auto-linked)Gemma
FamilyPart of the Gemma family
OrganizationGoogle DeepMind, Google Research
Organization typeBig-tech lab
CountryUnited Kingdom, United States
Model classDrug-discovery LLM (instruction-tuned)
DomainSmall molecules (also: Text, LLMs & agents)
First release
Basis: Developer blog announcing the open weights
Note: The model hub repository creation date of 2025-03-21 is not used as first release under this round's rule; the paper v1 is 2025-04-08 source
Other milestones
  • First public evidence: · Official blog source
  • Preprint v1: · Preprint v1 source
  • Weights release: · All three sizes released with the announcement, subject to the developer's terms source
  • Official announcement: · Official blog announcing open weights source
Latest update · maintenance only
Basis: Last modification date of the three model hub repositories
That date is a repository modification, not a new model version; no later TxGemma version was found as of 2026-11-16 source
The most recent sourced event, not a claim that nothing changed since.
StatusMaintained only · No new version within 12 months, only repository modifications
OpennessCode: Partly open
Weights: Research use only
Weights are open subject to the developer's health-AI terms; the released version was trained only on commercially licensed data, unlike the paper version
Benchmarks reported1 counted (self 1)
Wet-lab validationNone recorded
Unverified fields† open_status.weights_license, open_status.code_license
As of2026-11-16

What it does

Rewrites the tasks of a public therapeutics dataset library as instruction data and fine-tunes an open general base so that one model covers classification, regression and generation across small-molecule and protein therapeutic entities; the predict variant gives structured output, the chat variant natural-language explanation, and the paper adds an agent form. It is the open successor to the same organisation's earlier closed drug LLM and is often used as a comparator in papers such as Enchant and ether0.

Benchmarks reported

Every row is one pairing from the survey's relation file; a score belongs to the benchmark page's ranking, where it is shown with metric, eval set, source type and date. 1 of 1 pairings are counted.

BenchmarkRelationEval set / protocolSource typeNote and evidence
04 TDC (ADMET + DTI)Developer report, official protocol—Self-reported66 TDC tasks, recommended splits · evidence

Other evaluations (outside the 44)

  • Agent-form task demonstrations (Own task set) Self-reported: The paper wires the model to tools for multi-step task demonstrations, author-defined and outside the benchmarks on this site

Developers report: Developers report that a single open-weights model covers most tasks of that dataset library, matching or beating specialist models on most of them.

Independent evidence: A company ran it as a comparator on its own kinase task set: a third-party evaluation, though not on an official split.

Training data

Group in Figure 7: Public data only.

ClassValueDetail
a. Public databasesYes66 tasks and over 15 million data points from a public therapeutics dataset library source
b. Private or licensed dataNoThe public weights use only commercially licensed datasets; the paper version also used non-commercial ones source
c. Own wet-lab data used for trainingNo source
d. Wet lab used only for validationNo source

Wet-lab validation

None. The developers report no wet-lab validation.

Background

  • Organization: The developer is a large technology company's AI research unit, founded in London in 2010, acquired in 2014 and merged with the company's other research team in 2023-04; this model was released through its health-AI developer programme as internal research.
  • Funding: Not disclosed
  • Products and partnerships: Its predecessor was the same organisation's closed drug LLM on a different base; the base here is the company's open general model; related structural-biology work is AlphaFold 3.
  • Related models: AlphaFold 3, Enchant v1/v2, ether0

Reading notes

  1. Its selling point is covering many tasks in one model, not topping any single task.
  2. The paper keeps separate counts against specialist and against general-purpose comparators; the two must not be merged when quoted.
  3. The paper version and the public weights were trained on different data, so reproducing the paper with the open weights may fall short.
  4. Training and evaluation come from the same dataset library, so transfer to real projects is unverified.

Caveats

  1. First release follows the official blog; the model hub's repository creation date is not used as evidence.
  2. 2025-04-10 is a repository modification date and must not be presented as a new version.
  3. The third-party kinase task set is not that library's official split, so its figures cannot be listed beside official-protocol ones.

Sources

  1. Paper preprint (preprint)
  2. Model card (official page)
  3. Launch blog (blog) · 2025-03-25
  4. Model hub repository (model hub)
  5. Third-party evaluation post (blog)

Corrections

2 corrections
FieldWasNowBasisFound
latest_updateNo later update (N/A)Repository change on 2025-04-10, kind recorded as maintenance only, not a new versionsource2026-11-16
first_release2025-03 (from memory, not sourced)2025-03-25, based on the official release blogsource2026-11-16

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