Comprehensive Chinese Evaluation on C-Eval
89AccuracyQwen3
Evaluation Results
| Method | Links | |
|---|---|---|
| Qwen3Model=Qwen3, Number of Parameters=32B, Privacy Protocol=Plaintext2026.03 | 89 | |
| NTele-R1-32B2025.08 | 88.82 | |
| R1-DistillModel=R1-Distill, Number of Parameters=32B, Privacy Protocol=Plaintext2026.03 | 88.34 | |
| DS-32B2025.08 | 87.98 | |
| AloePriModel=R1-Distill, Number of Parameters=32B, Privacy Protocol=AloePri2026.03 | 87.84 | |
| DS-V3.1-Terminus (no_think)Model=DS-V3.1-Terminus (no_think), Number of Parameters=671B, Privacy Protocol=Plaintext2026.03 | 87.67 | |
| AloePriModel=Qwen3, Number of Parameters=32B, Privacy Protocol=AloePri2026.03 | 87.64 | |
| Qwen3Model=Qwen3, Number of Parameters=14B, Privacy Protocol=Plaintext2026.03 | 87.35 | |
| AloePriModel=Qwen3, Number of Parameters=14B, Privacy Protocol=AloePri2026.03 | 87.12 | |
| Qwen3-MoE-InstructModel=Qwen3-MoE-Instruct, Number of Parameters=30B-A3B, Privacy Protocol=Plaintext2026.03 | 86.97 | |
| AloePriModel=Qwen3-MoE-Instruct, Number of Parameters=30B-A3B, Privacy Protocol=AloePri2026.03 | 86.79 | |
| R1-DistillModel=R1-Distill, Number of Parameters=14B, Privacy Protocol=Plaintext2026.03 | 86.18 | |
| AloePriModel=R1-Distill, Number of Parameters=14B, Privacy Protocol=AloePri2026.03 | 85.93 | |
| AloePriModel=DS-V3.1-Terminus (no_think), Number of Parameters=671B, Privacy Protocol=AloePri2026.03 | 85.65 | |
| Qwen2.5 (base)Params=7B, Tokens=18T, Complexity Type=Quadratic, Evaluation Framework=vLLM, Evaluation Method=perplexity-based2025.09 | 81.68 | |
| SpikingBrain-76B (76B-A12B)Params=12B/76B, Tokens=+160B, Complexity Type=Hybrid, Evaluation Framework=vLLM, Evaluation Method=perplexity-based2025.09 | 78.89 | |
| FPModel=LLADA-1.5, Precision=Full Precision2026.06 | 70.05 | |
| FPModel=LLADA, Precision=Full Precision2026.06 | 69.54 | |
| MoE + L_ERCNumber of Parameters=15B, Activated Parameters=700M, Loss Function=ERC loss2025.12 | 69 | |
| STaR-QuantModel=LLADA-1.5, Quantization=W8A82026.06 | 68.96 | |
| STaR-QuantModel=LLADA, Quantization=W8A82026.06 | 67.87 | |
| MoENumber of Parameters=15B, Activated Parameters=700M2025.12 | 67.5 | |
| FPModel=DREAM, Precision=Full Precision2026.06 | 64.89 | |
| STaR-QuantModel=DREAM, Quantization=W8A82026.06 | 62.89 | |
| Gemma2-27BParams=27B, Tokens=13T, Complexity Type=Quadratic, Evaluation Framework=vLLM, Evaluation Method=perplexity-based2025.09 | 61.02 | |
| Mixtral-8*7B (47B-A13B)Params=13B/47B, Complexity Type=Quadratic, Evaluation Framework=vLLM, Evaluation Method=perplexity-based2025.09 | 54.39 | |
| Llama3Model=Llama3, Number of Parameters=8B, Privacy Protocol=Plaintext2026.03 | 50.16 | |
| LLama2-70bParams=70B, Tokens=2T, Complexity Type=Quadratic, Evaluation Framework=vLLM, Evaluation Method=perplexity-based2025.09 | 49.26 | |
| Jamba (52B-A12B)Params=12B/52B, Complexity Type=Hybrid, Evaluation Framework=vLLM, Evaluation Method=perplexity-based2025.09 | 48.94 | |
| AloePriModel=Llama3, Number of Parameters=8B, Privacy Protocol=AloePri2026.03 | 44.55 |