PanoptiveLabs
Longitudinal evaluation of AI agentsHow we checked the grader ↗

ClinOpsBench: Trial Oversight v1.0

Agent trajectories

Follow a run from source inspection to its final assessment.

One representative run per model is published, the run with the median composite. Remaining runs are available on request. Request access ↗

How the representative run is selected

Runs are ordered by saved composite, then run number, then run ID; the middle run is selected.

15 runs · 15 models
Recorded trajectory

Follow the agent's work

627 saved calls · 190 lifecycle events

Completed episode
E0Call 1

Browse sources

tool:list_artifacts
1 / 627

2027-01-04 · Evidence update

Tool execution

Input

Recorded
Empty object · 0 fields

Output

Recorded
data
1 fields
artifacts
21 items
21 records
artifact iddisplay namemedia typepage countrow countvisible
demographics
demographics.csv
text/csv
null (recorded value)20true
ecog
ecog.csv
text/csv
null (recorded value)21true
edc_audit_trail
edc_audit_trail.csv
text/csv
null (recorded value)16true
eligibility_criteria
eligibility_criteria.csv
text/csv
null (recorded value)160true
informed_consent
informed_consent.csv
text/csv
null (recorded value)25true
local_labs
local_labs.csv
text/csv
null (recorded value)7true
medical_history
medical_history.csv
text/csv
null (recorded value)68true
prior_anticancer_therapy
prior_anticancer_therapy.csv
text/csv
null (recorded value)43true
requirements
requirements.csv
text/csv
null (recorded value)34true
treatment_exposure
treatment_exposure.csv
text/csv
null (recorded value)1true
tumor_assessments
tumor_assessments.csv
text/csv
null (recorded value)20true
visits
visits.csv
text/csv
null (recorded value)22true
vital_signs
vital_signs.csv
text/csv
null (recorded value)3true
EV-CMP-001
EV-CMP-001.pdf
application/pdf
5null (recorded value)true
EV-MVR-GUIDE-001
EV-MVR-GUIDE-001.pdf
application/pdf
3null (recorded value)true
EV-PROTOCOL-A9
EV-PROTOCOL-A9.pdf
application/pdf
102null (recorded value)true
EV-REG-MATRIX-001
EV-REG-MATRIX-001.pdf
application/pdf
3null (recorded value)true
EV-SOP-MON-001
EV-SOP-MON-001.pdf
application/pdf
3null (recorded value)true
EV-SOP-RBQM-001
EV-SOP-RBQM-001.pdf
application/pdf
2null (recorded value)true
EV-SOW-001
EV-SOW-001.pdf
application/pdf
2null (recorded value)true
ok
true
Inspect displayed JSON