List your company

Knowledge Reasoning on K-Cross (test)

41.5AccuracyCMA
29.22832.41435.638.786May 28, 2026

Evaluation results

MethodLinks
CMABase model=Qwen2.5-1.5B2026.0541.5
PSO-MergingBase model=Qwen2.5-1.5B2026.0541.4
Model SwarmBase model=Qwen2.5-1.5B2026.0541.1
PSO-MergingBackbone=Qwen2.5-1.5B2026.0539.3
Single BestBase model=Qwen2.5-1.5B2026.0539.2
Single BestBackbone=Qwen2.5-1.5B2026.0539.2
CMABackbone=Qwen2.5-1.5B2026.0539.2
Model SwarmBackbone=Qwen2.5-1.5B2026.0539.2
Task ArithmeticBase model=Qwen2.5-1.5B2026.0539.1
TABackbone=Qwen2.5-1.5B2026.0539.1
DAREBase model=Qwen2.5-1.5B2026.0539
DAREBackbone=Qwen2.5-1.5B2026.0539
EvoGMBase model=Qwen2.5-1.5B2026.0538.8
EvoGMBackbone=Qwen2.5-1.5B2026.0537.9
TIESBase model=Qwen2.5-1.5B2026.0534.2
TIESBackbone=Qwen2.5-1.5B2026.0534.2
Model SoupBase model=Qwen2.5-1.5B2026.0533
Model SoupBackbone=Qwen2.5-1.5B2026.0533
MTLBase model=Qwen2.5-1.5B2026.0532.7
MTLBackbone=Qwen2.5-1.5B2026.0532.7
BaseBase model=Qwen2.5-1.5B2026.0529.7
BaseBackbone=Qwen2.5-1.5B2026.0529.7