APEX-Agents · Management Consulting
World132_SF_Task03
APEX-Agents task World132_SF_Task03 in AI Agents for Financial Statement Analysis. Compare dual-harness agent runs across models — rubric criteria, scores, and public traces.
Task prompt
What the agent was asked to do
What is the impact on the forecasted 2027 EBITDA in Brazil if we reallocate half of a 20% savings on ingredients towards increasing our marketing spend? We want to achieve a 10% bump in sales. If the result in Brazil is better than +4.583%, run the same scenario in Japan's forecast. Round all inputs, intermediate and final calculations to 4 decimal places. Return all your final outputs to me here as a message.
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 | 2/2 | Pass | Share pagePublic trace |
| Gemini 3.1 Pro | dual | 2/2 | Pass | Share pagePublic trace |
| GPT-5.4 | dual | 0/2 | Fail | Share pagePublic trace |
| GPT-5.4 mini | dual | 1/2 | Fail | Share pagePublic trace |
| GPT-5.4 nano | dual | 2/2 | Pass | Share pagePublic trace |
Grading rubric
Criteria and grader verdict (showcase run)
States that the forecast 2027 EBITDA in Brazil is -$5.7250M
PassEvidence: The response table lists Brazil “Revised EBITDA” as “-5.7250,” and the Brazil calculation also states “Revised EBITDA | -5.7250.” Assessment: The criterion asks whether it states the forecast 2027 EBITDA in Brazil is -$5.7250M; pass because the response clearly reports Brazil’s scenario EBITDA as -5.7250 million dollars.
States that the forecast 2027 EBITDA in Japan is -$4.5580M
PassEvidence: The response table lists Japan “Revised EBITDA” as “-4.5580,” and the Japan calculation also states “Revised EBITDA | -4.5580.” Assessment: The criterion asks whether it states the forecast 2027 EBITDA in Japan is -$4.5580M; pass because the response clearly reports Japan’s scenario EBITDA as -4.5580 million dollars.