Instruction Following Evaluation on IFEval
86.69IFEval ScoreDS-V3.1-Terminus (no_think)
Evaluation Results
| Method | Links | |
|---|---|---|
| DS-V3.1-Terminus (no_think)Model=DS-V3.1-Terminus (no_think), Number of Parameters=671B, Privacy Protocol=Plaintext2026.03 | 86.69 | |
| Qwen3Model=Qwen3, Number of Parameters=14B, Privacy Protocol=Plaintext2026.03 | 86.14 | |
| AloePriModel=Qwen3, Number of Parameters=14B, Privacy Protocol=AloePri2026.03 | 85.4 | |
| AloePriModel=DS-V3.1-Terminus (no_think), Number of Parameters=671B, Privacy Protocol=AloePri2026.03 | 84.66 | |
| Qwen3Model=Qwen3, Number of Parameters=32B, Privacy Protocol=Plaintext2026.03 | 84.29 | |
| ZeroRL + ACTGUIDE-RL2026.05 | 82.99 | |
| ZeroRL2026.05 | 81.33 | |
| AloePriModel=Qwen3, Number of Parameters=32B, Privacy Protocol=AloePri2026.03 | 81.15 | |
| Qwen3-MoE-InstructModel=Qwen3-MoE-Instruct, Number of Parameters=30B-A3B, Privacy Protocol=Plaintext2026.03 | 80.41 | |
| AloePriModel=Qwen3-MoE-Instruct, Number of Parameters=30B-A3B, Privacy Protocol=AloePri2026.03 | 78.56 | |
| SFT2026.05 | 77.82 | |
| SFT + ACTGUIDE-RL2026.05 | 77.43 | |
| SFT + RL2026.05 | 76.34 | |
| R1-DistillModel=R1-Distill, Number of Parameters=32B, Privacy Protocol=Plaintext2026.03 | 73.94 | |
| AloePriModel=R1-Distill, Number of Parameters=32B, Privacy Protocol=AloePri2026.03 | 73.57 | |
| R1-DistillModel=R1-Distill, Number of Parameters=14B, Privacy Protocol=Plaintext2026.03 | 72.83 | |
| AloePriModel=R1-Distill, Number of Parameters=14B, Privacy Protocol=AloePri2026.03 | 72.27 | |
| CrispEditData=ZsRE 10K2026.02 | 69.9 | |
| LLaMA-3-8B-InstructData=ZsRE 10K2026.02 | 69.3 | |
| UltraEditData=ZsRE 10K2026.02 | 68.9 | |
| Llama3Model=Llama3, Number of Parameters=8B, Privacy Protocol=Plaintext2026.03 | 68.58 | |
| AloePriModel=Llama3, Number of Parameters=8B, Privacy Protocol=AloePri2026.03 | 67.84 | |
| LocBF-FTData=ZsRE 10K2026.02 | 67.6 | |
| + Safety 4k + Feature synth 4kSFT training data category=Trained on general SFT data + safety SFT data + SAE synthetic data2026.05 | 53.23 | |
| + Safety 8kSFT training data category=Trained on general SFT data + safety SFT data2026.05 | 53.05 | |
| Alpaca 50kSFT training data category=Trained on general SFT data only2026.05 | 51.94 | |
| + Safety 4k + Random synth 4kSFT training data category=Trained on general SFT data + safety SFT data + SAE synthetic data2026.05 | 48.98 | |
| + Safety 120kSFT training data category=Trained on general SFT data + safety SFT data2026.05 | 48.06 | |
| + Safety 40kSFT training data category=Trained on general SFT data + safety SFT data2026.05 | 47.5 | |
| + Safety 200kSFT training data category=Trained on general SFT data + safety SFT data2026.05 | 47.5 | |
| AlphaEditData=ZsRE 10K2026.02 | 20.9 | |
| Adam-NSCLData=ZsRE 10K2026.02 | 14.8 |