BenchSci vs Edison Scientific (2026)
BenchSci and Edison Scientific both give researchers AI systems that investigate literature and scientific data. BenchSci's EMET emphasizes an evidence workspace connected to organizational knowledge; Edison's Kosmos undertakes extended research runs combining literature search, executable analysis and hypothesis generation. EMET uses a biomedical knowledge graph with literature, experimental databases and internal research. Organizations can connect laboratory systems and build a private graph, supported by scientific staff and customized workflows. Current examples include tuberculosis research and kidney experiment planning. BenchSci claims EMET does not hallucinate, though its own terms acknowledge that outputs can be incorrect. Kosmos takes an objective and dataset and can work unattended for up to 12 hours. Its report links statements to code or source literature, leaving scientists to assess the resulting conclusions. A technical report describes seven discoveries and a 79.4% statement accuracy assessment. Edison also acknowledges that longer runs can pursue statistically significant but scientifically irrelevant findings. BenchSci fits groups wanting a continuing research environment built around shared experimental knowledge and enterprise support. Edison fits researchers assigning a bounded objective and dataset to an extended computational investigation, with review after the run. Its enterprise offer adds deployment in the customer's environment and custom world models.
- Research depends on internal laboratory knowledge and literature together.
- Teams want a private biomedical knowledge graph.
- Scientific support should accompany configured enterprise workflows.
- Eligible academic users need limited free research access.
- A research objective includes a dataset for executable analysis.
- Researchers can review results after an extended unattended run.
- Conclusions need traceability to analysis code and literature.
- Enterprise deployment should run in the customer's environment.
This comparison is published by AI Health Index, an independent research platform that compares healthcare AI vendors objectively. BenchSci and Edison Scientific are each graded against the same capability taxonomy, from each vendor's own public materials and the regulatory record, under the AI Health Index verification standard. No vendor pays for placement, and no vendor has reviewed this page. How this evidence is graded
Plain facts
| Fact | BenchSci | Edison Scientific |
|---|---|---|
| Primary category | Drug Discovery AI | Drug Discovery AI |
| Founded | 2015 | 2025 |
| Headquarters | Toronto, Canada | San Francisco, California, United States |
| Website | benchsci.com | edisonscientific.com |
Side by Side
Select any grade to read the note behind it.
Each vendor in one paragraph
Written to be quoted whole. Each paragraph opens the AI Health Index profile of the vendor: what the product is, who buys it and what it does.
BenchSci develops EMET, an AI research environment for scientists in biopharma and academia. The platform combines scientific literature, experimental databases and internal research with a biomedical knowledge graph. Researchers ask questions in ordinary language, and agents select tools, search evidence and assemble analyses for work such as target discovery, experimental planning and drug safety research.
Source: AI Health Index, October 2026
Edison Scientific is the commercial spinout of the nonprofit research lab FutureHouse, formed in November 2025 to run its AI scientist platform. Its flagship agent, Kosmos, takes a research objective and a dataset and runs for up to 12 hours, reading literature, executing analysis code and generating hypotheses through a structured world model, then returns a report in which every statement is cited to code or a source.
Source: AI Health Index, October 2026
Questions buyers ask
How do BenchSci and Edison Scientific work with the EHR?
For BenchSci, the July 2026 internal data release names live connections to Benchling, Veeva, electronic laboratory notebooks, D360 and GDB, alongside document uploads and a private knowledge graph option. It distinguishes reading existing systems from ingesting material into that graph.
With Edison Scientific, the direction of any data flow, and where it runs in production, are not described either. From the AI Health Index, drawn from each vendor's own published materials and the regulatory record, verified October 15, 2026. No vendor pays for placement.
How are BenchSci and Edison Scientific priced and sold?
In BenchSci's case, enterprise access is custom priced and starts with a free pilot. The enterprise package includes broader content access, internal data connections, advanced reasoning, team controls, onboarding and dedicated support.
For Edison Scientific, the Kosmos announcement of 5 November 2025 launched it at $200 a run (200 credits at $1 a credit), with free usage for academics and founding subscriptions that lock in the $1 credit price, and says the price "ultimately will probably be higher"; users sign up on the platform without contacting sales, and enterprise deployments are quoted separately. From the AI Health Index, drawn from each vendor's own published materials and the regulatory record, verified October 15, 2026. No vendor pays for placement.
How do BenchSci and Edison Scientific handle patient data?
With BenchSci, the privacy notice separately requires consent before identifiable personal information trains external models, while allowing certain internal research uses. Routine retention follows purpose and legal need rather than a fixed deletion schedule; records associated with safety incidents have a minimum five year period.
The privacy policy (effective 30 October 2025), which covers the platform's inputs and outputs, says Edison does "not and will not intentionally train on any user's uploaded data, such as any scientific data that may be uploaded". It does train on what it calls trainable data, meaning agent trajectories and user interaction data, unless the user opts out in the account. From the AI Health Index, drawn from each vendor's own published materials and the regulatory record, verified October 15, 2026. No vendor pays for placement.
Which settings and specialties do BenchSci and Edison Scientific cover?
For BenchSci, the Cape Town tuberculosis example and the kidney experiment case support actual use in two research settings. Sanofi's ASCEND license extends to its global preclinical organization across therapeutic areas, although that agreement does not validate every newer EMET workflow.
With Edison Scientific, the technical report's discoveries span metabolomics, materials science, neuroscience and statistical genetics, and the home page positions Kosmos across the drug development lifecycle from target identification through clinical development and filing, for biopharma organizations and academic researchers. From the AI Health Index, drawn from each vendor's own published materials and the regulatory record, verified October 15, 2026. No vendor pays for placement.
Related comparisons
Other published head to head assessments involving these vendors or their closest peers. The full set for this category is on the Drug Discovery AI page.
BenchSci is based in Toronto, Canada. Edison Scientific is a San Francisco commercial spinout of FutureHouse. Edison's $200 per run figure is a launch price, not a current tier schedule. Both enterprise offers have negotiated prices.