List your company

Bank Statement Question Answering on MortarBench 1.0 (test)

83.6Overall F1Gemini 3.1 Pro + CRIT
5562.42569.8577.275Jun 17, 2026

Evaluation results

MethodLinks
Gemini 3.1 Pro + CRITBackbone Model=Gemini 3.1 Pro, CRIT Agent Integration=true2026.0683.680.5977796.729.313.36
Gemini 3.1 ProBackbone Model=Gemini 3.1 Pro, CRIT Agent Integration=false2026.0680.877.19772.296.736.71.100
GPT-5.5 + CRITBackbone Model=GPT-5.5, CRIT Agent Integration=true2026.0680.277.393.573.195311.27.714
GPT-5.5Backbone Model=GPT-5.5, CRIT Agent Integration=false2026.0679.976.894.672.393.334.71.91.33
Claude Sonnet 4.6 + CRITBackbone Model=Claude Sonnet 4.6, CRIT Agent Integration=true2026.0662.359.436.972.69034.31.22.75
Claude Sonnet 4.6Backbone Model=Claude Sonnet 4.6, CRIT Agent Integration=false2026.0656.151.436.3638548.72.51.75