List your company

Reasoning on PolyMath 8 Seen Languages

50PolyMath ScoreMGDA
10.4820.743141.26Jun 4, 2026

Evaluation results

MethodLinks
MGDABackbone=Qwen3-8B-Base2026.0650
CAGradBackbone=Qwen3-8B-Base2026.0649
PCGradBackbone=Qwen3-8B-Base2026.0648.4
V. SFTBackbone=Qwen3-8B-Base2026.0646.8
CAGradBackbone=Qwen3-4B-Base2026.0644.9
MGDABackbone=Qwen3-4B-Base2026.0644.7
PCGradBackbone=Qwen3-4B-Base2026.0644.6
BaseBackbone=Qwen3-8B-Base2026.0644.5
V. SFTBackbone=Qwen3-4B-Base2026.0643.4
BaseBackbone=Qwen3-4B-Base2026.0639.4
CAGradBackbone=Llama-3.1-8B2026.0622.9
MGDABackbone=Llama-3.1-8B2026.0622.4
PCGradBackbone=Llama-3.1-8B2026.0622.4
MGDABackbone=Meta-Llama-3-8B2026.0621.6
PCGradBackbone=Meta-Llama-3-8B2026.0621.4
V. SFTBackbone=Llama-3.1-8B2026.0620.3
CAGradBackbone=Meta-Llama-3-8B2026.0620.2
V. SFTBackbone=Meta-Llama-3-8B2026.0618.3
BaseBackbone=Llama-3.1-8B2026.0613.2
BaseBackbone=Meta-Llama-3-8B2026.0612