Co-Scientist
Google DeepMind's multi-agent AI system that generates, debates, and evolves novel scientific hypotheses — validated in a Nature paper.
Updated 2026-06-06
Yes. Co-Scientist is free to use.
Listed pricing: Free (experimental access via registration).
Overview
Co-Scientist is Google DeepMind's multi-agent research system, built on Gemini, that acts as a virtual scientific collaborator. Rather than just searching or summarizing papers, it generates novel hypotheses, then uses a tournament-style debate among specialized AI agents to refine, challenge, and rank those hypotheses before presenting the strongest ones to the researcher. A Nature paper published May 19, 2026 validated the approach across drug discovery and biomedical research, with lab partners confirming that Co-Scientist-generated hypotheses led to genuinely useful experimental directions.
The architecture is what sets it apart from standard research tools. Multiple specialist agents — a generation agent, a critique agent, a ranking agent, and a meta-review agent — work in parallel, debating each hypothesis against the existing literature and known constraints. This isn't just RAG over papers; it's an attempt at automated scientific reasoning, where weak hypotheses get eliminated through structured adversarial review before a human ever sees them.
Access is currently experimental and free via registration at labs.google/science. Google is positioning this alongside other Gemini-for-science tools, and enterprise previews are available for institutional research teams. The obvious limitation is scope: it's strongest in biomedical and drug discovery domains where the training data and validation partnerships are concentrated. Whether it generalizes well to physics, materials science, or social sciences remains to be seen.
Is Co-Scientist free?
Yes. Co-Scientist is free to use.
What the free tier covers: Full experimental access available via registration at labs.google/science.
Listed pricing: Free (experimental access via registration).
Pricing on this page has not been re-verified. The entry was last edited on , and no separate pricing check has been run since. Treat the figures as a record of what was published then and confirm on the official site.
What the free tier leaves out
Read straight off the plan list below. Vendors move features between tiers, so check the current split before you pay.
- Enterprise Preview Contact Google Institutional access for research teams. Custom onboarding, priority access, and dedicated support.
Co-Scientist pricing
| Plan | Price | What's included |
|---|---|---|
| Experimental Access | Free | Register at labs.google/science for individual researcher access. Includes hypothesis generation, multi-agent debate, and literature synthesis. |
| Enterprise Preview | Contact Google | Institutional access for research teams. Custom onboarding, priority access, and dedicated support. |
Register at labs.google/science for individual researcher access. Includes hypothesis generation, multi-agent debate, and literature synthesis.
Institutional access for research teams. Custom onboarding, priority access, and dedicated support.
One plan carries no published number: Enterprise Preview, recorded as “Contact Google”. Nothing is estimated in its place. The vendor's own pricing page is the only source for those figures.
Pricing on this page has not been re-verified. The entry was last edited on , and no separate pricing check has been run since. Treat the figures as a record of what was published then and confirm on the official site.
Is Co-Scientist worth it?
Worth it for Biomedical and drug-discovery researchers who want AI-generated, debated hypotheses grounded in literature before designing experiments.
You can test that on the free tier before paying anything. The recorded trade-offs are listed below, and any one of them can settle the question on its own.
The 8.2/10 AI Score is an editorial read of published capability, price and shipping pace. Nobody here has hands-on hours with Co-Scientist. How we verify.
Worth it if
The strengths recorded against this entry.
- Genuinely novel approach — multi-agent debate produces higher-quality hypotheses than single-model generation
- Nature-published validation with real lab partnerships confirming practical value in drug discovery
- Free experimental access with no paywall — rare for a tool this ambitious
- Goes beyond search and summarization to actual hypothesis generation and scientific reasoning
Not worth it if
Any one of these blocks your use case.
- Strongest in biomedical and drug discovery — generalization to other scientific domains is unproven
- Experimental status means features, access, and availability could change without notice
- No API or programmatic access — interaction is through Google's interface only
- Requires registration and approval, so access isn't instant
What sets Co-Scientist apart
- Tournament-style debate among generator, critic, ranker, and meta-review agents
- Validated in a Nature paper with real lab partnerships in drug discovery
- Free experimental access via registration, no paywall
- No API — interaction only through Google's interface
Key features
Hypothesis Generation
Generates novel scientific hypotheses from a researcher's goals and constraints, going beyond literature search to propose testable ideas grounded in existing knowledge.
Multi-Agent Debate
Specialized agents — generators, critics, rankers, and meta-reviewers — debate each hypothesis in a tournament-style system, eliminating weak proposals through structured adversarial review.
Literature Synthesis
Draws on scientific literature to ground hypotheses in existing research, identifying gaps, contradictions, and opportunities across published work.
Experimental Design
Suggests experimental approaches to test generated hypotheses, including methodologies and key variables to control for.
How it compares
| Tool | Best for | Pricing | Score |
|---|---|---|---|
| Co-Scientist | Biomedical and drug-discovery researchers who want AI-generated, debated hypotheses grounded in literature before designing experiments. | Free (experimental access via registration) | 8.2/10 |
| Perplexity AI vs Perplexity AI → | Knowledge workers and researchers who want cited, synthesized answers instead of a list of links to click through. | Freemium | 9.4/10 |
| Gemini 3.8 Flash | — | $0.75 per million input tokens, $3.75 per million output tokens | 9.2/10 |
| NotebookLM | Students, researchers, and professionals who need answers grounded strictly in the specific documents they upload. | Free | 9.1/10 |
Compare head-to-head
Comparison explorer Put Co-Scientist up against any three tools Opens with Co-Scientist already loaded. Add up to three more from the full index and read pricing, features, pros and cons in one table. →Related reading
GPT-6 Astra Expands to Paid Users After September Launch
OpenAI released GPT-6 Astra on September 3 to Daybreak customers before wider rollout, with pricing at $10 input and $50 output per million tokens on
Astra Token Rates Against Sol Predecessor
OpenAI documentation lists GPT-6 Astra at ten dollars input and fifty dollars output per million tokens, compared to four and twenty for GPT-5.6 Sol.
GPT-6 Astra Is Live for Limited Orgs — Not GA Yet
Help Center and release notes still say limited orgs / not GA / gradual. OpenAI's X post claims Pro, Enterprise, and Business Premium are live in Work/Codex plus API — note that separately. API still lists $10/$50.
Ready to try Co-Scientist?
Head to the official site to start with Co-Scientist — pricing and plans are listed above.
Visit Co-Scientist