Multi-hop Question Answering on MusiQue answerable setting
38.98ConcisenessGemini-3-pro
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| Gemini-3-proSize=UNK2026.01 | 38.98 | 59.21 | 94.1 | |
| GPT-o3Size=UNK2026.01 | 37.3 | 60.83 | 96.48 | |
| Qwen3Size=235B-A22B2026.01 | 35.25 | 60.94 | 95.62 | |
| DeepSeek-R1Size=671B2026.01 | 34.33 | 55.82 | 91.12 | |
| GPT-4oSize=UNK2026.01 | 32.54 | 57.1 | 93.66 | |
| Qwen3Size=30B-A3B2026.01 | 32.41 | 53.69 | 85.06 | |
| Qwen3Size=14B2026.01 | 31.78 | 48.81 | 85.6 | |
| Qwen3Size=8B2026.01 | 31.33 | 52.06 | 86.71 | |
| DeepSeek-V3.2Size=671B2026.01 | 30.34 | 59.5 | 86.33 | |
| Qwen3Size=4B2026.01 | 29.83 | 46.88 | 84.4 | |
| LLaMA-3.1Size=70B2026.01 | 27.72 | 51.92 | 86.06 | |
| GPT-5.1Size=UNK2026.01 | 27.27 | 60.11 | 87.79 | |
| Claude-4.5Size=UNK2026.01 | 27.13 | 63.88 | 89.66 | |
| Qwen3Size=1.7B2026.01 | 21.56 | 30.39 | 73.05 | |
| SelfCiteSize=8B2026.01 | 19.02 | 35.16 | 75.59 | |
| LongCiteSize=8B2026.01 | 17.55 | 32.73 | 76.04 | |
| LLaMA-3.1Size=8B2026.01 | 12.35 | 28.55 | 65.96 | |
| Qwen3Size=0.6B2026.01 | 6.92 | 11.32 | 58.12 | |
| FRONTSize=13B2026.01 | 6.34 | 13.7 | 61.19 | |
| LLaMA-3.2Size=3B2026.01 | 3.46 | 7.43 | 55.42 | |
| LLaMA-3.2Size=1B2026.01 | 2.37 | 3.82 | 26.03 |