List your company

Tool Use on ToolSandBox (test)

69.88Overall Score235B-A22B-Instruct
38.149646.387354.62562.8627Jun 5, 2026

Evaluation results

MethodLinks
235B-A22B-InstructGroup=Open-source Baselines2026.0669.8862.4662.6863.8559.4862.6857.6390.9873.8473.4474.9870.7574.8570.9974.7374.79
Traj.-level TRUSTBase Model=Qwen3-8B-Base2026.0668.2868.6659.763.156.9361.5852.0490.773.0176.4174.9469.1666.7468.7771.2671.24
AUQBase Model=Qwen3-8B-Thinking2026.0664.0856.4656.6857.8553.4856.6851.6387.9867.8467.4468.9864.7568.8564.9968.7368.79
30B-A3B-InstructGroup=Open-source Baselines2026.0661.5762.0860.0762.8854.0962.3457.2458.4564.6862.1262.2561.7263.6164.9463.5463.55
Traj.-level CM2Base Model=Qwen3-8B-Base2026.0661.2170.5758.5564.2251.9356.6453.755.660.1265.0465.0265.0361.861.364.1464.51
8B-ThinkingGroup=Open-source Baselines2026.0659.7464.3244.7950.1644.0661.5535.7694.9359.0562.9565.1360.7462.3561.0865.2364
SAGEBase Model=Qwen3-8B-Thinking2026.0659.7254.8551.5653.9547.5659.4143.6883.4859.765.7260.2758.2962.9766.6162.2265.52
Turn-level TRUSTBase Model=Qwen3-4B-Thinking2026.0656.3551.1846.2652.8337.6855.5942.2794.4457.4258.5359.0558.6857.6656.958.0258.77
AUQBase Model=Qwen3-4B-Thinking2026.0655.4748.2347.9249.244.7951.5941.7184.8756.1258.0960.5556.9860.4159.0355.5156.99
4B-ThinkingGroup=Open-source Baselines2026.0652.8957.6838.0943.1538.1853.0328.7190.8258.7154.7553.2152.6153.8253.6456.460.54
SAGEBase Model=Qwen3-4B-Thinking2026.0652.5448.2246.9349.1442.0454.3840.9971.0650.1456.0354.3351.660.4651.4955.955.45
Turn-level GRPOBase Model=Qwen3-4B-Thinking2026.0639.3746.9221.1528.6719.0423.3719.0595.9843.36—40.9338.8644.2837.941.8243.71