List your company
ResearchFormerly WizWand

General Utility Evaluation on MT_Bench

82.7Agreement RateC2PO
31.53244.81658.171.384Dec 29, 2025Jan 23, 2026Feb 17, 2026Mar 14, 2026Apr 8, 2026May 3, 2026May 28, 2026
Updated 5mo ago

Evaluation results

MethodLinks
C2POBackbone=DeepSeek-R1-Distill-Qwen-14B2025.1282.7
RubricHubJudge Model=Claude Sonnet 4, Approach Type=Training-free Approach2026.0581.72
Human-crafted (existing) RubricsJudge Model=Claude Sonnet 4, Rubric Granularity=Dataset-specific2026.0581.62
Training-free Rubric GenerationJudge Model=Claude Sonnet 4, Rubric Granularity=Instance-specific2026.0581.62
DnA-EvalJudge Model=Claude Sonnet 4, Approach Type=Training-free Approach2026.0581.51
Training-free Rubric GenerationJudge Model=Llama 3.1 70B, Rubric Granularity=Instance-specific2026.0580.98
Human-crafted (existing) RubricsJudge Model=Llama 3.1 70B, Rubric Granularity=Dataset-specific2026.0580.87
Training-free Rubric GenerationJudge Model=Qwen3 14B, Rubric Granularity=Instance-specific2026.0580.55
Training-free Rubric GenerationJudge Model=Claude Sonnet 4, Rubric Granularity=Dataset-specific2026.0580.55
RubricHubJudge Model=Llama 3.1 70B, Approach Type=Training-free Approach2026.0580.34
Training-free Rubric GenerationJudge Model=Qwen3 14B, Rubric Granularity=Dataset-specific2026.0580.02
Human-crafted (existing) RubricsJudge Model=Qwen3 14B, Rubric Granularity=Dataset-specific2026.0579.91
Training-free Rubric GenerationJudge Model=Llama 3.1 70B, Rubric Granularity=Dataset-specific2026.0579.91
CPOBackbone=LLaMA-2-13B-Chat2025.1279.9
C2POBackbone=LLaMA-2-13B-Chat2025.1279.9
RubricHubJudge Model=Qwen3 14B, Approach Type=Training-free Approach2026.0579.76
DnA-EvalJudge Model=Qwen3 14B, Approach Type=Training-free Approach2026.0578.8
DnA-EvalJudge Model=Llama 3.1 70B, Approach Type=Training-free Approach2026.0577.15
CPOBackbone=DeepSeek-R1-Distill-Qwen-14B2025.1277.1
CheckEvalJudge Model=Qwen3 14B, Approach Type=Training-free Approach2026.0575.24
IPOBackbone=DeepSeek-R1-Distill-Qwen-14B2025.1274.2
Prometheus 8x7BApproach Type=Fine-tuned Model, Rubric Granularity=Dataset-specific2026.0574.18
CheckEvalJudge Model=Claude Sonnet 4, Approach Type=Training-free Approach2026.0573.06
CheckEvalJudge Model=Llama 3.1 70B, Approach Type=Training-free Approach2026.0571.04
IPOBackbone=LLaMA-2-13B-Chat2025.1267.3
GRPOBackbone=DeepSeek-R1-Distill-Qwen-14B2025.1267
BCOBackbone=DeepSeek-R1-Distill-Qwen-14B2025.1264.7
DPOBackbone=DeepSeek-R1-Distill-Qwen-14B2025.1264.2
BCOBackbone=LLaMA-2-13B-Chat2025.1260.3
DPOBackbone=LLaMA-2-13B-Chat2025.1248.5
FRBackbone=LLaMA-2-13B-Chat2025.1245.4
FRBackbone=DeepSeek-R1-Distill-Qwen-14B2025.1239.7
GRPOBackbone=LLaMA-2-13B-Chat2025.1233.5