Raycaster / evalsBack to AI Agents for Infrastructure Finance

APEX-Agents · Investment Banking

World227_SK_Task04

1/1Pass

APEX-Agents task World227_SK_Task04 in AI Agents for Infrastructure Finance. Compare dual-harness agent runs across models — rubric criteria, scores, and public traces.

AI Agents for Infrastructure FinanceInvestment Banking World 227Dual harnessGrader: rubric
task_13a13a4b392745dc8b3d1bbfabe06016
Investment Banking World 227
message_in_console
4 models · dual config

Task prompt

What the agent was asked to do

You have the MFC model. Calculate the Total Debt / EBITDA in FY2030 and report it in multiple, rounded to one decimal place. Print your answer to me here. Reference the "Refinancing (Option A +B)” tab for interim calculations under Option A, Case 1 of the model “toggle.” Develop one scenario with the following specifications: Increase the Refinancing spread from 4.0% to 6.0%. Hold revenue growth constant at 6.0% per year from 2026 to 2030.

Published trajectories

Agent runs on this task

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

ModelHarnessScoreResultLinks
GPT-5.5showcasedual1/1Pass
Gemini 3.1 Produal1/1Pass
GPT-5.4 minidual1/1Pass
GPT-5.4 nanodual1/1Pass

Grading rubric

Criteria and grader verdict (showcase run)

  1. States the leverage ratio (Total Debt / EBITDA) in FY2030 is 2.5x

    Pass

    Evidence: <TEXT_RESPONSE> says "2.5x". Assessment: Criterion requires stating the FY2030 Total Debt / EBITDA leverage ratio is 2.5x; the response exactly states this, so it passes.