List your company

Knowledge-based Reasoning on Filtered CL-Bench (test)

34.96Accuracy (DK)No Memory
25.662428.076230.4932.9038Jun 3, 2026

Evaluation results

MethodLinks
No MemoryModel=Qwen3-4B2026.0634.961025.7136.3630.45
No MemoryModel=Qwen3-8B2026.0634.966.6735.7134.8532.18
TMEMModel=Qwen3-4B2026.0634.9616.6734.2934.8532.87
TMEMModel=Qwen3-8B2026.0634.151042.8634.8533.91
MemAgentModel=Qwen3-8B2026.0633.336.6738.5734.8532.18
AMEMModel=Qwen3-8B2026.0631.71103030.328.72
AMEMModel=Qwen3-4B2026.0628.461025.7133.3326.99
MemAgentModel=Qwen3-4B2026.0626.02103033.3326.99