| Model |
Provider |
Agent |
Easy Score |
Easy Rwd |
Medium Score |
Medium Rwd |
Hard Score |
Hard Rwd |
Very Hard Score |
Very Hard Rwd |
Time |
One-Shot Baseline — Grader Score by Difficulty Tier
Easy
Medium
Hard
Very Hard
Multi-Step Agent — Grader Score by Difficulty Tier
Easy
Medium
Hard
Very Hard
Cumulative Reward — Hard Task
Cumulative Reward — Very Hard Task
Total Reward Across All Tasks