Overview
| Name | Co-Scientist · aliases: AI co-scientist, Google Co-Scientist |
|---|---|
| Full name | AI co-scientist |
| Ambiguous names (not auto-linked) | Co-Scientist, co-scientist |
| Family | Not a family entry. A single system; the 2025 preprint and 2026 journal version are versions within one record |
| Organization | Google DeepMind, Google Research |
| Organization type | Big-tech lab |
| Country | United Kingdom, United States |
| Model class | Agent (hypothesis generation) |
| Domain | Text, LLMs & agents |
| First release | · paper revision Basis: arXiv preprint v1 (official blog the same period) source |
| Other milestones | |
| Latest update | · paper revision Basis: arXiv v2 arXiv v2 posted (scope of revision not inspected), after the 2026-05-19 Nature paper and opening to individual researchers through Gemini for Science source The most recent sourced event, not a claim that nothing changed since. |
| Status | Product only, no public artifact · Closed; offered through an experimental tool and a cloud preview |
| Openness | Code: Closed Weights: Closed Closed; individual researchers apply via Gemini for Science and an enterprise version is in Google Cloud preview |
| Benchmarks reported | 0 counted (unconfirmed 6) |
| Wet-lab validation | 1 recorded campaign |
| Unverified fields | † wetlab_validation[0], release_milestones.official_announcement, training_data.a_public |
| As of | 2026-11-17 |
What it does
A Gemini-based multi-agent system in which generation, reflection, ranking, evolution and meta-review agents iterate on hypotheses, ranking them in an Elo tournament of pairwise comparisons and simulated debates, and output research proposals; checks call web search and public databases such as ChEMBL and UniProt. Unlike Kosmos or Biomni, it outputs ranked hypotheses rather than running data analysis.
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. 0 of 6 pairings are counted.
| Benchmark | Relation | Eval set / protocol | Source type | Note and evidence |
|---|---|---|---|---|
| 14 PubMedQA | Unconfirmed (not counted) | — | Unconfirmed source | No report found in official materials †disputed · evidence |
| 18 LAB-Bench / LABBench2 | Unconfirmed (not counted) | — | Unconfirmed source | No report found in official materials †disputed · evidence |
| 23 BioASQ | Unconfirmed (not counted) | — | Unconfirmed source | No report by it was found in the official material. †disputed · evidence |
| 29 BixBench | Unconfirmed (not counted) | — | Unconfirmed source | No report found in official materials †disputed · evidence |
| 41 BioML-bench | Unconfirmed (not counted) | — | Unconfirmed source | No report on this benchmark found in official materials. †disputed · evidence |
| 42 BioMysteryBench | Unconfirmed (not counted) | — | Unconfirmed source | No report on this benchmark found in official materials. †disputed · evidence |
Other evaluations (outside the 44)
- Site text and agent benchmarks (none disclosed) Self-reported: Checked LAB-Bench, BixBench, BioASQ, PubMedQA, BioML-bench, BioMysteryBench: not reported in the paper or official blog
- Elo vs GPQA correctness (Public benchmark not among the 44) Self-reported: Uses GPQA Diamond to check whether system Elo tracks answer correctness; the Elo is self-assessed with no independent ground truth
- Expert evaluation (user or partner report) Self-reported: A small expert panel rated novelty and impact; the developers describe the sample as small
- Elo tournament (Own task set) Self-reported: Self-play ranking improves with test-time compute; no independent ground truth
Developers report: a multi-agent Gemini system that generates and ranks hypotheses, with Elo tracking quality and some hypotheses confirmed in partner labs.
Independent evidence: Press reports describe a UK university team trying it on a bacterial-evolution question, and 2025 press criticism noted a lack of independent evaluation; no systematic third-party study was found.
Training data
Group in Figure 7: Public data + wet lab for validation only.
| Class | Value | Detail |
|---|---|---|
| a. Public databases | Not disclosed | Base Gemini training data undisclosed; ChEMBL, UniProt and similar are inference-time tools source |
| b. Private or licensed data | Not disclosed | |
| c. Own wet-lab data used for training | No | The system is not trained on wet-lab data |
| d. Wet lab used only for validation | Yes | Some hypotheses were tested in partner labs; the developers give no aggregate counts 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.
Denominator basis: No aggregate counts; individual cases such as liver-fibrosis targets, AML drug repurposing and cellular senescence, each run by a partner lab
Qualitative cases only; must not be converted into hit rates or quantitative validation
Executor: Run by external partner labs; not independent replication
Background
- Organization: Google DeepMind was founded in London in 2010, acquired by Google in 2014 and merged with Google Brain in 2023-04; this system was built with Google Research, Google Cloud and Google Labs.
- Funding: A wholly owned internal research unit of Alphabet with no separate funding.
- Products and partnerships: Enterprise preview partners include Daiichi Sankyo, Bayer Crop Science and US national laboratories; academic cases include Stanford, MIT, Edinburgh, Cambridge and Calico; the developers cite researchers at over 100 institutions.
- Related models: Kosmos / Edison Analysis, Biomni, Robin, MedGemma / Med-Gemini, VERID
Reading notes
- It has no standard site benchmark score and cannot be compared directly with Kosmos or Biomni.
- Partner-lab results are single cases without stated controls or replicates and cannot be read as efficacy or hit rates.
- The Elo-GPQA correlation is self-assessed and comes from the 2025 preprint.
- The 2026 claim of planning experiments has no detail in official text, so it is best read as a hypothesis-generation system.
Caveats
- Wet lab was done by partner labs with no aggregate counts; treated as unverified.
- Closed, so external evaluation is not possible.
- The scope of the arXiv v2 revision was not inspected.
Sources
- Nature paper (journal)
- arXiv preprint (preprint)
- DeepMind official blog (2026-05-19) (company blog)
- Google Research 2025 blog (company blog)
- Gemini for Science launch blog (company blog)
- Nature paper registry record (journal)
Corrections
2 corrections
Part of the AI for Bio benchmarks survey · all 112 models · open in the explorer. Record as of 2026-11-17.