List your company

Multimodal Agent Performance on HR-Bench 4k

81.9AccuracyBaseline
75.86877.4347980.566Jun 2, 2026

Evaluation results

MethodLinks
BaselineBackbone=235B-FP8, Training regime=None2026.0681.9
ToolGateBackbone=235B-FP8, Training regime=Cross-domain2026.0680.5
ToolGateBackbone=30B, Training regime=In-domain2026.0677.3
ToolGateBackbone=30B, Training regime=Cross-domain2026.0676.3
BaselineBackbone=30B, Training regime=None2026.0676.1