Token Usage per Task
Average input, cached input, output, and reasoning tokens per task.
ExploitHunter.app
Eval Explorer
Loading eval results from the generated explorer dataset.
Eval Categories
Model Groups
Providers
Time per Task
Average active eval runtime. Lower is better.
Cost Distribution per Task
Adjusted P5, P25, P50, P75, and P95 cost spread for each model.
Composite Model Index
Category-balanced normalized score across selected eval categories. Higher is better.
Category Profile
Score shape across eval families for the selected result set.
Score vs. Cost
Paid-model value view: stronger models with more completed tasks per dollar sit toward the upper right.
Run Logs
Collapsed model and run groups with normalized rows for audit and reruns.
| Model family | Eval | Task | Score | Cost | Time | Tokens | Calls / Cap |
|---|