Comparison on Loong Set 3: 100K–200K Tokens
57.84LLM ScoreDisco-RAG
Evaluation results
| Method | Links | ||
|---|---|---|---|
| Disco-RAGBase Model=Llama-3.3-70B2026.01 | 57.84 | 28 | |
| StructRAGCondition=SOTA Results2026.01 | 57.74 | 35 | |
| Disco-RAGBase Model=Qwen2.5-72B2026.01 | 56.89 | 19 | |
| Disco-RAGBase Model=Llama-3.1-8B2026.01 | 55.8 | 11 | |
| RQ-RAGCondition=SOTA Results2026.01 | 44.62 | 0 | |
| Llama-3.3-70BCondition=Standard RAG2026.01 | 43.7 | 6 | |
| Qwen2.5-72BCondition=Standard RAG2026.01 | 41.83 | 4 | |
| Llama-3.3-70BCondition=Full Context2026.01 | 41.11 | 14 | |
| Llama-3.1-8BCondition=Standard RAG2026.01 | 40.24 | 3 | |
| Qwen2.5-72BCondition=Full Context2026.01 | 40.13 | 13 | |
| Llama-3.1-8BCondition=Full Context2026.01 | 37.43 | 12 | |
| GraphRAGCondition=SOTA Results2026.01 | 27.4 | 0 |