APEX-Agents · Law
LawWorld433_ANB_02
APEX-Agents task LawWorld433_ANB_02 in AI Agents for Employment Law Analysis. Compare dual-harness agent runs across models — rubric criteria, scores, and public traces.
Task prompt
What the agent was asked to do
Jennifer Love was the nurse assigned to Mr. Buchanan's floor, and was later designated as the corporate designee of Magnolia Gardens Inc during depositions. She is called by Magnolia Gardens Inc., to testify in the wrongful death case against Magnolia Gardens Inc. Can Jennifer testify as to statements made by Tom Buchanan? Write out your reply to me in here with everything I requested.
Published trajectories
Agent runs on this task
Curated dual-harness runs (parsed + original sandbox). Best scored run per model.
| Model | Harness | Score | Result | Links |
|---|---|---|---|---|
| GPT-5.5showcase | dual | 4/4 | Pass | Share pagePublic trace |
| fireworks models Kimi K2 | dual | 0/4 | Fail | Share pagePublic trace |
| Gemini 3 Flash | dual | 3/4 | Fail | Share pagePublic trace |
| Gemini 3.1 Pro | dual | 0/4 | Fail | Share pagePublic trace |
| GPT-5.4 | dual | 1/4 | Fail | Share pagePublic trace |
| GPT-5.4 mini | dual | 0/4 | Fail | Share pagePublic trace |
| GPT-5.4 nano | dual | 2/4 | Fail | Share pagePublic trace |
Grading rubric
Criteria and grader verdict (showcase run)
States No, Jennifer Love cannot testify regarding statements made by Tom Buchanan
PassEvidence: TEXT_RESPONSE says, “Short answer: likely no” and “she should not be permitted…to testify…about statements allegedly made by Tom Buchanan.” Assessment: The criterion requires stating no, Jennifer cannot testify regarding Tom’s statements. Pass; the response clearly states she likely cannot, subject to exceptions.
States that the Dead Man's Act statute precludes a person testifying on behalf of an interested party from testifying to statements made by a deceased individual
PassEvidence: TEXT_RESPONSE states “The strongest objection is New York’s Dead Man’s Statute, CPLR 4519” and explains it bars “a party or interested witness” testifying “against the…administrator…about a personal transaction or communication with the deceased person.” Assessment: The criterion requires stating the Dead Man’s Act precludes a person testifying on behalf of an interested party from testifying to statements by a deceased individual. Pass; this is directly conveyed.
States that Magnolia Gardens is a party in interest in the wrongful death suit
PassEvidence: TEXT_RESPONSE says this is “the wrongful death case against Magnolia Gardens Inc.” and that Jennifer would be “testifying for Magnolia, against Daisy Buchanan as administrator/survivor.” It also refers to “Magnolia’s behalf” and “Magnolia…as corporate designee/party representative.” Assessment: The criterion requires stating Magnolia Gardens is a party in interest in the wrongful death suit. Pass; while it does not use the exact phrase “party in interest,” it identifies Magnolia as defendant/party whose behalf the testimony serves.
States that Jennifer Love testimony would be on behalf of Magonia Gardens, an interested party
PassEvidence: TEXT_RESPONSE states “if Magnolia calls Jennifer as its corporate designee or representative witness” and “Jennifer would be testifying for Magnolia,” plus “given the facts…she was designated Magnolia’s corporate designee and is being called by Magnolia.” Assessment: The criterion requires stating Jennifer Love’s testimony would be on behalf of Magnolia Gardens, an interested party. Pass; the response explicitly says she would testify for/on behalf of Magnolia.