Alignment on IFEval
86.3IFEval ScoreQwen2.5-VL-72B
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| Qwen2.5-VL-72BModel Type=Instruct, Parameters=72B2025.02 | 86.3 | — | |
| Llama-3.1-405BModel Type=Instruct, Parameters=405B2025.02 | 86 | — | |
| Qwen2.5-72BModel Type=Instruct, Parameters=72B2025.02 | 84.1 | — | |
| Llama-3.1-70BModel Type=Instruct, Parameters=70B2025.02 | 83.6 | — | |
| Qwen2-72B-InstructType=Instruction-tuned, Protocol=strict-prompt2024.07 | 77.6 | — | |
| Qwen2-72BModel Type=Instruct, Parameters=72B2025.02 | 77.6 | — | |
| Llama-3-70B-InstructType=Instruction-tuned, Protocol=strict-prompt2024.07 | 77.3 | — | |
| Mixtral-8x22B-InstructType=Instruction-tuned, Protocol=strict-prompt2024.07 | 67.1 | — | |
| Qwen1.5-110B-ChatType=Instruction-tuned, Protocol=strict-prompt2024.07 | 57.5 | — | |
| Qwen1.5-72B-ChatType=Instruction-tuned, Protocol=strict-prompt2024.07 | 55.8 | — | |
| CompassMax-V3mode=prompt strict2025.12 | — | 83.55 | |
| CompassMax-V3-Thinkingmode=prompt strict2025.12 | — | 85.4 | |
| DeepSeek-R1mode=prompt strict2025.12 | — | 84.84 | |
| Qwen3-0.6BModel Size=0.6B, Architecture=Standard, Prompt type=strict prompt, Few-shot setting=0-shot, Max tokens=327682026.02 | — | 60.6 | |
| Qwen3-0.6B-ALLMEMModel Size=0.6B, Architecture=ALLMEM, Sliding window length=2048, Attention sinks=128, Prompt type=strict prompt, Few-shot setting=0-shot, Max tokens=327682026.02 | — | 58.4 | |
| Qwen3-1.7BModel Size=1.7B, Architecture=Standard, Prompt type=strict prompt, Few-shot setting=0-shot, Max tokens=327682026.02 | — | 68.6 | |
| Qwen3-1.7B-ALLMEMModel Size=1.7B, Architecture=ALLMEM, Sliding window length=2048, Attention sinks=128, Prompt type=strict prompt, Few-shot setting=0-shot, Max tokens=327682026.02 | — | 69 |