Instruction Following on ComplexBench Out-of-Domain
79.8Overall ScoreQwen2.5-32B-Instruct-IF
Evaluation results
| Method | Links | |
|---|---|---|
| Qwen2.5-32B-Instruct-IFBase Model=Qwen-32B2025.10 | 79.8 | |
| Qwen2.5-32B-InstructBase Model=Qwen-32B2025.10 | 77.7 | |
| Distill-Qwen-14B-IFBase Model=Qwen-14B2025.10 | 76.9 | |
| QwQ-32BBase Model=Qwen-32B2025.10 | 73.3 | |
| Distill-Qwen-14BBase Model=Qwen-14B2025.10 | 72.7 | |
| GPT-4oBase Model=GPT2025.10 | 71.6 | |
| R1-0528-Qwen3-8BBase Model=Qwen-8B2025.10 | 71.1 | |
| Qwen2.5-7B-Instruct-IFBase Model=Qwen-7B2025.10 | 68.7 | |
| R1-0528-Qwen3-8BBase Model=Qwen-8B2025.10 | 68.5 | |
| Qwen2.5-7B-InstructBase Model=Qwen-7B2025.10 | 66.1 | |
| Llama-3.1-8B-Instruct-IFBase Model=LLaMA-8B2025.10 | 64.5 | |
| SPAR-8B-DPOBase Model=LLaMA-8B2025.10 | 63.8 | |
| Llama-3.1-8B-InstructBase Model=LLaMA-8B2025.10 | 63.6 | |
| Distill-Qwen-7B-IFBase Model=Qwen-7B2025.10 | 60.7 | |
| Crab-7B-DPOBase Model=Mistral-7B2025.10 | 59 | |
| Distill-Qwen-7BBase Model=Qwen-7B2025.10 | 55.2 | |
| VERIF-8BBase Model=LLaMA-8B2025.10 | 54.7 | |
| Qwen2.5-1.5B-Instruct-IFBase Model=Qwen-1.5B2025.10 | 48.6 | |
| Conifer-7B-DPOBase Model=Mistral-7B2025.10 | 48.1 | |
| Qwen2.5-1.5B-InstructBase Model=Qwen-1.5B2025.10 | 45.9 | |
| IR-1.5BBase Model=Qwen-1.5B2025.10 | 44.4 | |
| Distill-Qwen-1.5B-IFBase Model=Qwen-1.5B2025.10 | 43.3 | |
| Distill-Qwen-1.5BBase Model=Qwen-1.5B2025.10 | 39.8 |