A private 91-task knowledge-work evaluation spanning four multi-week professional projects and artifact-oriented analytical and presentation work.
Targets realistic, multi-step professional deliverables that short question-answer benchmarks miss. Separate analytical, presentation, rubric, and overall results could expose different knowledge-work failure modes if a compliant result source becomes available.
Ingestion is blocked until Artificial Analysis publishes a documented, versioned scored result API or artifact with exact system provenance and clear republication permission. AA-Briefcase Lite contains examples only and cannot substitute for the private scored suite.
Higher is better. No AgMoDB score is currently ingested; future overall, analytical, presentation, and rubric metrics must remain distinct and version-pinned.
Ingestion is blocked until Artificial Analysis publishes a documented, versioned scored result API or artifact with exact system provenance and clear republication permission.
No scores available for this benchmark yet.