Model               Mode              Phase1  Phase2  Phase3  Phase4  Overall  Tools
------------------------------------------------------------------------------------
Claude Sonnet       arise               0.60    0.73    1.00    0.80     0.78      2
Claude Sonnet       no tools            0.60    0.13    1.00    0.80     0.63      0
GPT-4o-mini         arise               0.20    0.67    0.93    0.47     0.57     21
GPT-4o-mini         no tools            0.33    0.07    0.93    0.60     0.48      0
GPT-4o-mini         fixed tools         0.13    0.53    0.87    0.40     0.48      0
