Raycaster/ Eval

APEX-Agents · Management Consulting

world130_HO_08

Best published1/2Fail

APEX-Agents task world130_HO_08 in AI Agents for Employment Law Analysis. Compare dual-harness agent runs across models, scores, and public traces.

AI Agents for Employment Law AnalysisManagement Consulting World 130Dual harnessGrader: rubric
task_1e19b26b91804fe386ef911e347df0f8
Management Consulting World 130
message_in_console
5 models · dual config

Task prompt

What the agent was asked to do

The client sent us employee wage data (attached), so we need to update our assumptions in the financial analysis section of the survey analysis report to display the updated annual productivity loss figures (in 000s). Find the average hourly salary of employee roles and use that to update the annual productivity loss estimate (rounded to 1 decimal place). Assume average hourly wages in the data are fully-loaded costs. Assume the following activities are non-productive: manual data entry, searching for data, and fixing errors. Report final answers here, written out in a short message.

Published trajectories

Agent runs on this task

Curated dual-harness runs (parsed + original sandbox). Best scored run per model.

ModelHarnessScoreResultLinks
Gemini 3.1 Produal1/2Fail
GPT-5.4 minidual1/2Fail
GPT-5.4 nanodual1/2Fail
GPT-5.4dual0/2Fail
GPT-5.5dual0/2Fail

Grading rubric

Rubric criteria

Runs are graded against these criteria. Open a run for model-specific verdicts.

  1. States the updated annual productivity loss is $50,494,000

  2. States the average fully-loaded hourly wage used to calculate the annual productivity loss is $28.90