APEX-Agents · Investment Banking

World225_km_06

6/6Pass

APEX-Agents task World225_km_06 in AI Agents for Tax Due Diligence. Compare dual-harness agent runs across models — rubric criteria, scores, and public traces.

AI Agents for Tax Due DiligenceInvestment Banking World 225Dual harnessGrader: rubric

Task IDtask_9b9cc4e93ba6412b893d88d6d59f0181

WorldInvestment Banking World 225

Expected outputmake_new_sheet

Published runs4 models · dual config

Task prompt

What the agent was asked to do

Model out the NPV of distributions shareholders would receive under REIT conversion. - There's the $1.2 billion E&P purge that gets taxed as ordinary income at 37% (E&P purge occurs at Year 0, annual distributions occur at end of Years 1–5). - There are ongoing REIT dividends in the $650-750mm range that qualify for the 20% Section 199A deduction. - Apply 199A only to annual REIT distributions; do not apply 199A to the E&P purge (tax purge at 37%). - Run sensitivities across 10%, and 12% discount rates over a 5-year horizon. Show discount rates vs distribution levels, populated with respective after-tax NPV per share. Then, show me the base case NPV as a percentage of both the strategic offer and current trading price. Round NPV per share to 2 decimal places. Round percentages to 1 decimal place. Create an xlsx that has all of your results.

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	6/6	Pass	Share page Public trace
Gemini 3.1 Pro	dual	6/6	Pass	Share page Public trace
GPT-5.4 mini	dual	6/6	Pass	Share page Public trace
GPT-5.4 nano	dual	6/6	Pass	Share page Public trace

Grading rubric

Criteria and grader verdict (showcase run)

States the NPV per share for the 10% discount rate and $650mm annual REIT distribution scenario is $8.74
Pass
Evidence: TEXT_RESPONSE sensitivity table shows row “10.0%” and column “$650mm annual distribution” as “$8.74”. Assessment: Criterion requires stating the NPV per share for the 10% discount rate and $650mm scenario is $8.74; pass.
States the NPV per share for the 10% discount rate and $750mm annual REIT distribution scenario is $9.68
Pass
Evidence: TEXT_RESPONSE sensitivity table shows row “10.0%” and column “$750mm annual distribution” as “$9.68”. Assessment: Criterion requires stating the NPV per share for the 10% discount rate and $750mm scenario is $9.68; pass.
States the NPV per share for the 12% discount rate and $650mm annual REIT distribution scenario is $8.44
Pass
Evidence: TEXT_RESPONSE sensitivity table shows row “12.0%” and column “$650mm annual distribution” as “$8.44”. Assessment: Criterion requires stating the NPV per share for the 12% discount rate and $650mm scenario is $8.44; pass.
States the NPV per share for the 12% discount rate and $750mm annual REIT distribution scenario is $9.33
Pass
Evidence: TEXT_RESPONSE sensitivity table shows row “12.0%” and column “$750mm annual distribution” as “$9.33”. Assessment: Criterion requires stating the NPV per share for the 12% discount rate and $750mm scenario is $9.33; pass.
States the Base Case NPV as % of Strategic Offer is 19.7%
Pass
Evidence: TEXT_RESPONSE base case table states “Base NPV as % of strategic offer | 19.7%”. Assessment: Criterion requires stating the Base Case NPV as % of Strategic Offer is 19.7%; pass.
States the Base Case NPV as % of Current Trading is 21.7%
Pass
Evidence: TEXT_RESPONSE base case table states “Base NPV as % of current trading price | 21.7%”. Assessment: Criterion requires stating the Base Case NPV as % of Current Trading is 21.7%; pass.