APEX-Agents · Law
Task yhzc9d1a
APEX-Agents task Task yhzc9d1a in AI Agents for SEC Disclosure Analysis. Compare dual-harness agent runs across models, scores, and public traces.
Task prompt
What the agent was asked to do
CrowdStrike sent the attached list of historic stock transactions (note the document lists transactions for both Class A and B stocks). We need to determine if any of these people would be a part of the class in the Plymouth County lawsuit, assuming no opt-outs. Look for the class requirements and the attached list of transactions, identify which individuals from the upload purchased Class A stock, and which of those did so during the Class Period to qualify for the class. Make a new sheet and list their names and each transaction for which the individual qualifies.
Published trajectories
Agent runs on this task
Curated dual-harness runs (parsed + original sandbox). Best scored run per model.
| Model | Harness | Score | Result | Links |
|---|---|---|---|---|
| fireworks models Kimi K2 | dual | 5/5 | Pass | Run detailsPublic trace |
| Gemini 3 Flash | dual | 5/5 | Pass | Run detailsPublic trace |
| Gemini 3.1 Pro | dual | 5/5 | Pass | Run detailsPublic trace |
| GPT-5.4 | dual | 5/5 | Pass | Run detailsPublic trace |
| GPT-5.4 mini | dual | 5/5 | Pass | Run detailsPublic trace |
| GPT-5.4 nano | dual | 5/5 | Pass | Run detailsPublic trace |
| GPT-5.5 | dual | 5/5 | Pass | Run detailsPublic trace |
Grading rubric
Rubric criteria
Runs are graded against these criteria. Open a run for model-specific verdicts.
States that Michael Lawson made a Class A Stock Purchase on 11/30/2023
States that Larry Stone made a Class A Stock Purchase on 4/3/2024
States that Geralt Tribeard made a Class A Stock Purchase on 7/28/2024
States that Susan Willow made a Class A Stock Purchase on 2/7/2024
States that Rayland Givens made a Class A Stock Purchase on 7/5/2024