Research and Analysis

Preparing data, grounding conclusions in evidence, and producing quantitative analyses from local datasets and documents.

40 tasks3 subcategories40 general tasksArtifact-based oracle scoring
Mean Score
64.55
general suite
Score ≥ 0.80
53.00%
R0.8
Full Credit
25.46%
R1
Averaged over all 40 tasks in this category, each evaluated under the 70 general model-harness configurations (7 harnesses × 10 models). Open a task to inspect its instruction, package files, and the trajectory of every run.