List your company

General Instruction Following Accuracy on IFBench

84.1AccuracyMulti-DPOP
56.33263.54170.7577.959Oct 10, 2025

Evaluation results

MethodLinks
Multi-DPOPBackbone=Qwen3-30B2025.1084.1
Base ModelBackbone=Qwen3-30B2025.1083.9
A-DPOBackbone=Qwen3-30B2025.1083.5
Multi-DPOPBackbone=Tulu-3.1-8B2025.1082.6
Base ModelBackbone=Tulu-3.1-8B2025.1082.4
A-DPOBackbone=Tulu-3.1-8B2025.1082.1
GRPOBackbone=Qwen3-30B2025.1081.3
Multi-DPOPBackbone=Qwen3-4B2025.1080.7
Base ModelBackbone=Qwen3-4B2025.1080.5
A-DPOBackbone=Qwen3-4B2025.1080.2
SFTBackbone=Qwen3-30B2025.1080.1
GRPOBackbone=Tulu-3.1-8B2025.1079.8
SFTBackbone=Tulu-3.1-8B2025.1078.6
GRPOBackbone=Qwen3-4B2025.1077.9
SFTBackbone=Qwen3-4B2025.1076.8
Multi-DPOPBackbone=Gemma-3-4B2025.1076.5
Base ModelBackbone=Gemma-3-4B2025.1076.3
A-DPOBackbone=Gemma-3-4B2025.1076
GRPOBackbone=Gemma-3-4B2025.1073.6
DPOBackbone=Qwen3-30B2025.1073.6
SFTBackbone=Gemma-3-4B2025.1072.5
DPOBackbone=Tulu-3.1-8B2025.1071.2
DPOBackbone=Qwen3-4B2025.1069.3
Multi-DPOPBackbone=GLM-4-9B2025.1069.2
Base ModelBackbone=GLM-4-9B2025.1069
A-DPOBackbone=GLM-4-9B2025.1068.7
GRPOBackbone=GLM-4-9B2025.1066.1
SFTBackbone=GLM-4-9B2025.1065.3
DPOBackbone=Gemma-3-4B2025.1064.8
DPOBackbone=GLM-4-9B2025.1057.4