List your company

Mathematical Reasoning on DAPO-Math OOD average

60.5Avg@16GEOSD
36.89243.02149.1555.279Jul 7, 2026

Evaluation results

MethodLinks
GEOSDBackbone=Qwen3-8B2026.0760.58.6
TrOPDBackbone=Qwen3-8B2026.0754.82.9
GEOSDBackbone=Olmo-3-7B-Think2026.0754.87.6
GRPOBackbone=Qwen3-8B2026.0753.61.7
TIPBackbone=Qwen3-8B2026.0753.31.4
SkewKLBackbone=Qwen3-8B2026.0752.70.8
π0Backbone=Qwen3-8B2026.0751.9—
JSDBackbone=Qwen3-8B2026.0751.80.1
TrOPDBackbone=Olmo-3-7B-Think2026.0750.43.2
TIPBackbone=Olmo-3-7B-Think2026.0750.12.9
GEOSDBackbone=DS-R1-Llama-8B2026.0749.45.7
SkewKLBackbone=Olmo-3-7B-Think2026.0747.90.7
GRPOBackbone=Olmo-3-7B-Think2026.0747.60.4
π0Backbone=Olmo-3-7B-Think2026.0747.2—
TIPBackbone=DS-R1-Llama-8B2026.0746.83.1
GRPOBackbone=DS-R1-Llama-8B2026.0746.22.5
JSDBackbone=Olmo-3-7B-Think2026.0746.11.1
RevKLBackbone=Qwen3-8B2026.0744.47.5
TrOPDBackbone=DS-R1-Llama-8B2026.0744.10.4
RevKLBackbone=Olmo-3-7B-Think2026.0743.93.3
FwdKLBackbone=Qwen3-8B2026.0743.88.1
π0Backbone=DS-R1-Llama-8B2026.0743.7—
JSDBackbone=DS-R1-Llama-8B2026.0742.90.8
SFTBackbone=Olmo-3-7B-Think2026.0742.84.4
FwdKLBackbone=Olmo-3-7B-Think2026.0742.64.6
SkewKLBackbone=DS-R1-Llama-8B2026.0741.52.2
RevKLBackbone=DS-R1-Llama-8B2026.0740.82.9
SFTBackbone=Qwen3-8B2026.0739.412.5
SFTBackbone=DS-R1-Llama-8B2026.0738.94.8
FwdKLBackbone=DS-R1-Llama-8B2026.0737.85.9