Citation Prediction Task 1 on CiteRAG 1.0 (test)
7.6Recall@20Gemini-2.5-Pro
Evaluation Results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| Gemini-2.5-ProRAG=R=102026.01 | 7.6 | 9.7 | 34.6 | 25.2 | 14.4 | 18.8 | |
| CitationGenerator-30BSFT=Yes, RAG=R=102026.01 | 7.6 | 12.2 | 36.7 | 29.5 | 19.3 | 21.9 | |
| CitationGenerator-30BSFT=Yes, RAG=R=52026.01 | 7.2 | 11.3 | 33.8 | 27.3 | 18.3 | 20.4 | |
| GPT-5RAG=R=102026.01 | 7.1 | 9.9 | 32.4 | 24.8 | 13.6 | 19.3 | |
| Gemini-2.5-ProRAG=R=52026.01 | 6.9 | 8.7 | 32.9 | 23.5 | 13.4 | 17 | |
| GPT-5RAG=R=52026.01 | 6.5 | 9.3 | 30.6 | 23.1 | 12.4 | 18.1 | |
| Claude-3.7RAG=R=102026.01 | 6.5 | 8.1 | 30 | 21.7 | 12.6 | 16.2 | |
| Claude-4-SonnetRAG=R=102026.01 | 6.4 | 8 | 29.6 | 21.3 | 12.2 | 15.7 | |
| GPT-5RAG=None2026.01 | 6.1 | 9 | 26.3 | 21.2 | 11.5 | 17.6 | |
| Kimi-K2RAG=R=102026.01 | 6.1 | 7.5 | 27.3 | 19.5 | 11.7 | 14.6 | |
| Deepseek-v3RAG=R=102026.01 | 6 | 7 | 27.8 | 19.3 | 11.1 | 13.7 | |
| Claude-3.7RAG=R=52026.01 | 5.7 | 7.1 | 27.5 | 19.6 | 11 | 14 | |
| GPT-o3RAG=R=102026.01 | 5.5 | 6.4 | 26.4 | 18.4 | 10.4 | 12.6 | |
| Gemini-2.5-ProRAG=None2026.01 | 5.5 | 6.9 | 27.2 | 19.5 | 10.8 | 13.8 | |
| Claude-3.5RAG=R=102026.01 | 5.4 | 6.2 | 26.4 | 18 | 10.3 | 12.1 | |
| Claude-4-SonnetRAG=R=52026.01 | 5.4 | 6.8 | 26.8 | 19.3 | 10 | 13.6 | |
| Kimi-K2RAG=R=52026.01 | 5.2 | 6.4 | 24.5 | 17.5 | 10 | 12.6 | |
| Deepseek-v3RAG=R=52026.01 | 5 | 5.8 | 24.8 | 17.1 | 9.6 | 11.4 | |
| Qwen3-Coder-480BRAG=R=102026.01 | 4.9 | 5.4 | 21.6 | 14.7 | 9.5 | 10.6 | |
| Gemini-2.0-FlashRAG=R=102026.01 | 4.8 | 5.8 | 23 | 16.2 | 9.2 | 11.4 | |
| GPT-o3RAG=R=52026.01 | 4.4 | 5.1 | 23 | 15.7 | 8.6 | 10.1 | |
| Claude-3.5RAG=R=52026.01 | 4.4 | 5 | 23 | 15.7 | 8.4 | 9.8 | |
| GPT-oss-120BRAG=R=102026.01 | 4.3 | 4.6 | 21.9 | 14.4 | 8.3 | 9 | |
| CitationGenerator-30BSFT=Yes, RAG=None2026.01 | 4.3 | 6.9 | 22.5 | 19.8 | 11.3 | 13 | |
| Claude-3.7RAG=None2026.01 | 4.2 | 5.4 | 19.7 | 14.4 | 8.1 | 10.7 | |
| Grok-3RAG=R=102026.01 | 4 | 4.2 | 19.2 | 12.2 | 7.7 | 7.9 | |
| Claude-4-SonnetRAG=None2026.01 | 3.8 | 5.1 | 18.7 | 13.9 | 7.6 | 10.3 | |
| Qwen3-Coder-480BRAG=R=52026.01 | 3.7 | 4.2 | 17.7 | 12.1 | 7.2 | 8.3 | |
| Kimi-K2RAG=None2026.01 | 3.6 | 4.6 | 17.7 | 12.7 | 7.1 | 9.1 | |
| Gemini-2.0-FlashRAG=R=52026.01 | 3.5 | 4.3 | 18 | 12.7 | 6.8 | 8.6 | |
| CitationGenerator-30BSFT=No, RAG=R=102026.01 | 3.1 | 3.5 | 15.3 | 10 | 5.9 | 6.3 | |
| GPT-o3RAG=None2026.01 | 2.9 | 3.4 | 16.6 | 11.3 | 5.9 | 6.8 | |
| Claude-3.5RAG=None2026.01 | 2.8 | 3.4 | 14 | 9.8 | 5.5 | 6.7 | |
| Deepseek-v3RAG=None2026.01 | 2.8 | 3.3 | 14.3 | 9.9 | 5.5 | 6.7 | |
| GPT-oss-120BRAG=R=52026.01 | 2.7 | 3.1 | 15.5 | 10.3 | 5.4 | 6 | |
| CitationGenerator-4BSFT=Yes, RAG=R=102026.01 | 2.7 | 3.9 | 12.9 | 9.8 | 5.3 | 7.4 | |
| Grok-3RAG=R=52026.01 | 2.4 | 2.5 | 13.4 | 8.7 | 4.7 | 4.9 | |
| Gemini-2.0-FlashRAG=None2026.01 | 2.3 | 3 | 11.5 | 8.4 | 4.5 | 6 | |
| CitationGenerator-4BSFT=Yes, RAG=R=52026.01 | 2.3 | 3.4 | 10.7 | 8.4 | 4.2 | 6.5 | |
| CitationGenerator-4BSFT=No, RAG=R=102026.01 | 2.1 | 2.3 | 12.2 | 7.8 | 4.1 | 4.5 | |
| CitationGenerator-30BSFT=No, RAG=R=52026.01 | 2.1 | 2.4 | 11.7 | 7.6 | 3.9 | 4.2 | |
| Qwen3-Coder-480BRAG=None2026.01 | 1.9 | 2.3 | 10 | 7.1 | 3.7 | 4.7 | |
| Grok-3RAG=None2026.01 | 1.6 | 1.7 | 9 | 6 | 3.1 | 3.4 | |
| GPT-oss-120BRAG=None2026.01 | 1.6 | 1.8 | 9.5 | 6.3 | 3.3 | 3.7 | |
| CitationGenerator-4BSFT=Yes, RAG=None2026.01 | 1.2 | 1.9 | 5.5 | 4.7 | 2.2 | 3.9 | |
| CitationGenerator-4BSFT=No, RAG=R=52026.01 | 1.1 | 1.2 | 5.6 | 3.4 | 2.4 | 2.7 | |
| CitationGenerator-30BSFT=No, RAG=None2026.01 | 1 | 1.1 | 5.7 | 3.8 | 2 | 2.3 | |
| CitationGenerator-4BSFT=No, RAG=None2026.01 | 0.6 | 0.7 | 2.9 | 2.1 | 1.3 | 1.7 |