Instruction-following clustering on REASONCLUSTER (Overall)
68.42V-measure (%)C1-Qwen-14B
Evaluation Results
| Method | Links | |
|---|---|---|
| C1-Qwen-14BModel Type=Our Model2026.03 | 68.42 | |
| C1-Qwen-7BModel Type=Our Model2026.03 | 66.54 | |
| o3Model Type=Reasoning Model2026.03 | 65.08 | |
| Gemini 2.5 ProModel Type=Reasoning Model2026.03 | 61.82 | |
| QwQ-32BModel Type=Reasoning Model2026.03 | 54.78 | |
| GPT-oss-120BModel Type=Reasoning Model2026.03 | 52.31 | |
| Distill-Llama-70BModel Type=Reasoning Model2026.03 | 45.12 | |
| DeepSeek-R1Model Type=Reasoning Model2026.03 | 44.06 | |
| GPT-4.1Model Type=General Model2026.03 | 43.51 | |
| GPT-4oModel Type=General Model2026.03 | 41.26 | |
| Distill-Qwen-32BModel Type=Reasoning Model2026.03 | 34.96 | |
| Llama-3.1-70B-InstructModel Type=General Model2026.03 | 31.55 | |
| Qwen2.5-72B-InstructModel Type=General Model2026.03 | 29.06 |