Instruction Following on ComplexBench (test)
79.8CSRCRGC
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| CRGCBase Model=Claude-3.7-Sonnet, Model Type=Closed-source2026.06 | 79.8 | 83.6 | 81 | |
| CRGCBase Model=GPT-4o, Model Type=Closed-source2026.06 | 78.6 | 82.5 | 79.8 | |
| CRGCBase Model=Gemini-2.5-Pro, Model Type=Closed-source2026.06 | 77.5 | 81.4 | 78.9 | |
| CRGCBase Model=Llama-3.1-72B, Model Type=Open-source2026.06 | 75.6 | 79.4 | 76.8 | |
| CRGCBase Model=Qwen2.5-72B, Model Type=Open-source2026.06 | 74.8 | 78.7 | 76.1 | |
| Self-Reflection (SR)Base Model=Claude-3.7-Sonnet, Model Type=Closed-source2026.06 | 74.1 | 78 | 75.5 | |
| Self-Reflection (SR)Base Model=GPT-4o, Model Type=Closed-source2026.06 | 73.2 | 77.1 | 74.6 | |
| Chain-of-Thoughts (CoT)Base Model=Claude-3.7-Sonnet, Model Type=Closed-source2026.06 | 72.3 | 76.2 | 73.7 | |
| Self-Reflection (SR)Base Model=Gemini-2.5-Pro, Model Type=Closed-source2026.06 | 72.3 | 76.3 | 73.8 | |
| CRGCBase Model=Mixtral-8x22B, Model Type=Open-source2026.06 | 72.2 | 76.1 | 73.5 | |
| Chain-of-Thoughts (CoT)Base Model=GPT-4o, Model Type=Closed-source2026.06 | 71.5 | 75.4 | 72.8 | |
| Chain-of-Thoughts (CoT)Base Model=Gemini-2.5-Pro, Model Type=Closed-source2026.06 | 70.6 | 74.5 | 72 | |
| Self-Reflection (SR)Base Model=Llama-3.1-72B, Model Type=Open-source2026.06 | 69.7 | 73.8 | 71 | |
| Self-Reflection (SR)Base Model=Qwen2.5-72B, Model Type=Open-source2026.06 | 68.9 | 73 | 70.3 | |
| Chain-of-Thoughts (CoT)Base Model=Llama-3.1-72B, Model Type=Open-source2026.06 | 67.8 | 71.9 | 69.1 | |
| Standard Prompting (SP)Base Model=Claude-3.7-Sonnet, Model Type=Closed-source2026.06 | 67.4 | 71.6 | 68.8 | |
| Chain-of-Thoughts (CoT)Base Model=Qwen2.5-72B, Model Type=Open-source2026.06 | 66.8 | 70.9 | 68.2 | |
| Standard Prompting (SP)Base Model=GPT-4o, Model Type=Closed-source2026.06 | 66.7 | 70.8 | 68.1 | |
| Self-Reflection (SR)Base Model=Mixtral-8x22B, Model Type=Open-source2026.06 | 66.4 | 70.5 | 67.7 | |
| Standard Prompting (SP)Base Model=Gemini-2.5-Pro, Model Type=Closed-source2026.06 | 65.8 | 70 | 67.2 | |
| Chain-of-Thoughts (CoT)Base Model=Mixtral-8x22B, Model Type=Open-source2026.06 | 64.5 | 68.6 | 65.8 | |
| Standard Prompting (SP)Base Model=Llama-3.1-72B, Model Type=Open-source2026.06 | 63.4 | 67.7 | 64.7 | |
| Standard Prompting (SP)Base Model=Qwen2.5-72B, Model Type=Open-source2026.06 | 62.5 | 66.9 | 63.9 | |
| Standard Prompting (SP)Base Model=Mixtral-8x22B, Model Type=Open-source2026.06 | 60.3 | 64.5 | 61.7 |