List your company
ResearchFormerly WizWand

General Knowledge Assessment on C-Eval

92.5AccuracyKimi-K2 Base
20.2238.98557.7576.515Jun 19, 2024Oct 5, 2024Jan 22, 2025May 11, 2025Aug 28, 2025Dec 15, 2025Apr 3, 2026
Updated 6mo ago

Evaluation results

MethodLinks
Kimi-K2 Base# Shots=5-shot, # Activated Params=32B, # Total Params=1043B2026.0192.5
Qwen3.5-35B-A3B#Token=8002026.0491.5
DeepSeek-V3.2 Exp Base# Shots=5-shot, # Activated Params=37B, # Total Params=671B2026.0191
DeepSeek-V3.1 Base# Shots=5-shot, # Activated Params=37B, # Total Params=671B2026.0190
Qwen3-Next-80B-A3B#Token=7002026.0489.2
JoyAI-LLM Flash#Token=6002026.0488.7
MiMo-V2-Flash Base# Shots=5-shot, # Activated Params=15B, # Total Params=309B2026.0187.9
Qwen3-30B-A3B#Token=8002026.0487.8
AUTOIF (Qwen2-72B-Instruct w/ Online DPO)Backbone Model=Qwen2-72B-Instruct, Supervision Model=Qwen2-72B, Alignment Strategy=Self-Alignment, Training Method=Online DPO2024.0684.9
Qwen2-72B-InstructBackbone Model=Qwen2-72B-Instruct2024.0683.8
PreBackbone=Qwen2.5-7B Instruct2025.1279.5
MMKEBackbone=Qwen2.5-7B Instruct, Consolidation=false2025.1279.3
MMKE+Con.Backbone=Qwen2.5-7B Instruct, Consolidation=true2025.1278.8
Qwen-1.5 14BModel Type=Teacher2024.0778.68
EtConBackbone=Qwen2.5-7B Instruct2025.1278.5
FT-M+Con.Backbone=Qwen2.5-7B Instruct, Consolidation=true2025.1277
GLM-4.7-Flash-T#Token=122002026.0477
AUTOIF (Qwen2-7B Online DPO)Backbone Model=Qwen2-7B, Supervision Model=Qwen2-72B, Alignment Strategy=Strong-to-Weak, Training Method=Online DPO2024.0676
FT-MBackbone=Qwen2.5-7B Instruct, Consolidation=false2025.1275.9
AUTOIF (Qwen2-7B Offline DPO)Backbone Model=Qwen2-7B, Supervision Model=Qwen2-72B, Alignment Strategy=Strong-to-Weak, Training Method=Offline DPO2024.0675.1
Qwen2-7BBackbone Model=Qwen2-7B2024.0674.4
AUTOIF (Qwen2-7B-SFT)Backbone Model=Qwen2-7B, Supervision Model=Qwen2-72B, Alignment Strategy=Strong-to-Weak, Training Method=SFT2024.0673.9
Qwen2-7B(ShareGPT)Backbone Model=Qwen2-7B2024.0670.2
DDKTeacher Model=Qwen-1.5 14B, Student Model=Qwen-1.5 4B2024.0768.57
KDTeacher Model=Qwen-1.5 14B, Student Model=Qwen-1.5 4B2024.0768.35
MiniLLMTeacher Model=Qwen-1.5 14B, Student Model=Qwen-1.5 4B2024.0768.2
CPTTeacher Model=Qwen-1.5 14B, Student Model=Qwen-1.5 4B2024.0768.05
Qwen-1.5 4BModel Type=Student2024.0767.6
Engram-40BShots=5-shot2026.0163.3
Engram-27BShots=5-shot2026.0162.7
AUTOIF (LLaMA-3-70B w/ Online DPO)Backbone Model=LLaMA3-70B-Instruct, Supervision Model=LLaMA3-70B, Alignment Strategy=Self-Alignment, Training Method=Online DPO2024.0661.6
LLaMA3-70B-InstructBackbone Model=LLaMA3-70B-Instruct2024.0660.2
MoE-27BShots=5-shot2026.0158
HY-1.8B-FP16Model Scale=1.8B, Precision=FP16, Instruction-tuned=true2026.0254.27
HY-1.8B-INT4Model Scale=1.8B, Precision=INT4, Instruction-tuned=true2026.0248.67
HY-1.8B-2BitModel Scale=1.8B, Precision=2Bit, Instruction-tuned=true2026.0247.6
Dense-4BShots=5-shot2026.0146.9
Mistral-7BBackbone Model=Mistral-7B2024.0638.2
AUTOIF (LLaMA3-8B Online DPO)Backbone Model=LLaMA3-8B, Supervision Model=LLaMA3-70B, Alignment Strategy=Strong-to-Weak, Training Method=Online DPO2024.0638.2
AUTOIF (LLaMA3-8B Offline DPO)Backbone Model=LLaMA3-8B, Supervision Model=LLaMA3-70B, Alignment Strategy=Strong-to-Weak, Training Method=Offline DPO2024.0636.2
HY-0.5B-FP16Model Scale=0.5B, Precision=FP16, Instruction-tuned=true2026.0235.98
LLaMA3-8B(ShareGPT)Backbone Model=LLaMA3-8B2024.0635.2
AUTOIF (LLaMA3-8B-SFT)Backbone Model=LLaMA3-8B, Supervision Model=LLaMA3-70B, Alignment Strategy=Strong-to-Weak, Training Method=SFT2024.0634.5
LLaMA3-8BBackbone Model=LLaMA3-8B2024.0624.2
ALPHABackbone=Qwen2.5-7B Instruct, Consolidation=false2025.1223
ALPHA+Con.Backbone=Qwen2.5-7B Instruct, Consolidation=true2025.1223