DecisionRAG CQA - Paragraph R@10: leaderboard
DecisionRAG v1 uses Jev (typesafe/jev-1.13), dense retrieval with Qwen3-Embedding-8B, BGE (bge-reranker-v2-m3), a Qwen3 8B reranker, and GPT 5 with low reasoning. Evaluation populations are MD2D with 4201 validation turns, CQA with 271 evidence-labeled development questions, MTRAG with 188 labeled IBM Cloud turns, and NQ with 500 seeded DAPR test questions. Main MTRAG gold labels are expanded using GPT 5 to judge the pooled top 5 documents; comparisons using the original official labels remain separate. Trained judges use Qwen3 Base at 1.7B, 4B, and 8B; their results are distinct from backbone language-model scores. All source-qualified harnesses and training variants remain separate. Metric: Accuracy (%).
Metric: Accuracy (%). Source: arxiv.org. Saturation forecast: Rough model projection: around 2026. 14 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | DecisionRAG study DecisionRAG (8B trained), 5 pages | 80.2 | |
| 2 | DecisionRAG study DecisionRAG (8B trained), 3 pages | 78.8 | |
| 3 | DecisionRAG study DecisionRAG (Jev) + dense candidates | 76.3 | |
| 4 | DecisionRAG study DecisionRAG (Jev), 5 pages | 73.2 | |
| 5 | DecisionRAG study DecisionRAG (Jev) | 71.9 | |
| 6 | DecisionRAG study Dense top-100 + Qwen3-Reranker | 67.2 | |
| 7 | DecisionRAG study Dense top-20 + GPT-5 | 64 | |
| 8 | DecisionRAG study Dense top-100 + BGE | 60.6 | |
| 9 | DecisionRAG study Dense | 54 | |
| 10 | DecisionRAG trained 8B / general 5pages | 45.5 |
Interactive version: theaggregate.ai/benchmark?slug=decisionrag-cqa-paragraph-r-10 · How It Works · Data refreshed daily, snapshot 2026-10-28.