APEX-Agents · Law
World431Task_EL_01
APEX-Agents task World431Task_EL_01 in AI Agents for Real Estate Lease Review. Compare dual-harness agent runs across models, scores, and public traces.
Task prompt
What the agent was asked to do
After execution, the lender's lawyers identified issues with the final estoppel certificate. Draft a new estoppel certificate that fixes the issues with the current "final" one. Create a new document and put your analysis in there.
Published trajectories
Agent runs on this task
Curated dual-harness runs (parsed + original sandbox). Best scored run per model.
| Model | Harness | Score | Result | Links |
|---|---|---|---|---|
| Gemini 3 Flash | dual | 6/6 | Pass | Run detailsPublic trace |
| GPT-5.4 mini | dual | 6/6 | Pass | Run detailsPublic trace |
| Gemini 3.1 Pro | dual | 4/6 | Fail | Run detailsPublic trace |
| GPT-5.4 nano | dual | 4/6 | Fail | Run detailsPublic trace |
| GPT-5.5 | dual | 4/6 | Fail | Run detailsPublic trace |
| GPT-5.4 | dual | 2/6 | Fail | Run detailsPublic trace |
| fireworks models Kimi K2 | dual | 1/6 | Fail | Run detailsPublic trace |
Grading rubric
Rubric criteria
Runs are graded against these criteria. Open a run for model-specific verdicts.
States that the Lessee is a signatory to the revised estoppel certificate
States that the Lessee certifies and warrants the facts listed in the revised estoppel certificate
States that there are no options to terminate the lease early in the revised estoppel certificate
States that the "Additional Rent" is the Lessee's proportionate share of operating expenses in the revised estoppel certificate
Replaces each instance of "Tenant" with "Lessee" in the revised estoppel certificate
Replaces each instance of "Landlord" with "Lessor" in the revised estoppel certificate