Jobs / Science and research
Can agents do science and research work?
Life, physical, and social science research and technical work.
- US jobs
- 1,473,280 (0.9% of employment)
- Typical wage
- $88K a year (employment-weighted mean of occupation medians)
- Occupations
- 66 in O*NET 31.0
- Evidence
- 1 direct benchmark, 4 also relevant, 0 context only
The answer today
Some tasksNo change since 10 Oct 2025.
Rule: Agents complete 20% or more of the tasks (or projects) on a direct benchmark.
The evidence that sets this step
- CORE-Bench by Princeton University: declared solved · Claude Code + Claude Opus 4.5 (after manual regrading), 3 Dec 2025
Next step: Most tasks needs 60% or more of the tasks (or projects) on two direct benchmarks from independent institutions.
How much of the work is measured
Less than 1% of the work in this family is measured by direct benchmarks. Little of the work is measured (less than 20%). On 10 Oct 2025: less than 1%.
Work = the O*NET 31.0 work activities of the family's occupations, weighted by how important each task is and by US employment. A benchmark measures an activity when its tasks make the agent do it and its grading checks it. Method.
| Work activity (O*NET) | Share of the work | Direct benchmark |
|---|---|---|
| Present research or technical information | 5.9% | Not measured |
| Research biological or ecological phenomena | 4.5% | Not measured |
| Analyze biological or chemical substances or related data | 4.3% | Not measured |
| Develop operational or technical procedures or standards | 2.9% | Not measured |
| Monitor operations to ensure adequate performance | 2.8% | Not measured |
| Direct scientific or technical activities | 2.6% | Not measured |
| Maintain operational records | 2.5% | Not measured |
| Develop research plans or methodologies | 2.3% | Not measured |
| Maintain current knowledge in area of expertise | 2.3% | Not measured |
| Inspect facilities or equipment | 2.1% | Not measured |
Our reading
Agents complete some science and research tasks, on one narrow benchmark. Princeton's HAL declared CORE-Bench solved in December 2025 after Claude Code with Claude Opus 4.5 scored 95.5% on 45 computational reproducibility tasks, after manual regrading of 8 tasks (77.78% under the automated grader). The tasks re-run published code that finishes in under 45 minutes, and scores depend heavily on the scaffold. No direct benchmark measures presenting research, researching biological or ecological phenomena, or analyzing substances and samples. Lab and field work is unmeasured. PaperBench, where the best official result is 26.0% (April 2025), maps here only in part and sets no step.
Evidence: CORE-Bench, PaperBench by OpenAI and Princeton University. Written 24 Sep 2026.
Benchmarks for this work
Direct evidence
Work benchmarks mapped primary to this family. They set the step when they report how many tasks agents complete.
- 95.5%Claude Code + Claude Opus 4.5 (after manual regrading) · 3 Dec 2025Solved
Also relevant
These benchmarks cover part of this family's work, or have no result yet. They do not set its step.
- 64.8%Claude Code + Opus 5.5 (max) · 6 Oct 2026Strong
Functions: Research replicationMedical image readingClinical records workData engineering and SQLComputer useEnterprise workflowsTerminal operationsFeature developmentML engineering
What we do not know
- Benchmarks measure tasks, not whole jobs. A job is a bundle of tasks, judgment, and relationships.
- Capability is not adoption. A task that agents can do in a benchmark may still be done by people at work.
- Most results come from single runs. Reliability over many runs is often unknown.
Occupations in this family
Sorted by US employment (BLS OEWS May 2025). O*NET sub-occupations share the employment figure of their SOC group. Work measured = the share of the occupation's O*NET work that direct benchmarks test.
How to cite
Credit the original work first: the benchmarks listed on this page, and their institutions.
Then, if you used this page:
Can Agents Work. "Can agents do science and research work?." https://canagentswork.com/jobs/science-research/ (accessed 2026-10-11). CC BY 4.0.@misc{caw-jobs-science-research,
title = {{Can agents do science and research work?}},
author = {{Can Agents Work}},
year = {2026},
howpublished = {\url{https://canagentswork.com/jobs/science-research/}},
note = {Accessed 2026-10-11. CC BY 4.0}
}