List your company

Multi-hop Question Answering on HotpotQA (EM, F1)

74.9F1CIRAG
8.44425.69742.9560.203Dec 20, 2025Dec 29, 2025Jan 8, 2026Jan 18, 2026Jan 27, 2026Feb 6, 2026Feb 16, 2026

Evaluation results

MethodLinks
CIRAGBase LLM=Qwen2.5-max2026.0174.960.9
MetaRAGBase LLM=Qwen2.5-max2026.0174.660.8
DualRAGBase LLM=Qwen2.5-max2026.0173.357.8
KiRAGBase LLM=Qwen2.5-max2026.0173.259
IRCOTBase LLM=Qwen2.5-max2026.0172.858.1
GraphAnchorBackbone=Qwen2.5-14B-Instruct2026.0169.6457.44
GraphAnchorBackbone=Qwen3-32B2026.0169.3455.85
CIRAGBase LLM=Qwen2.5-7B2026.0167.152.5
NativeRAGBase LLM=Qwen2.5-max2026.0166.952.4
GraphAnchorBackbone=Qwen2.5-7B-Instruct2026.0166.0353.8
GraphAnchorBackbone=Llama3.1-8B-Instruct2026.0164.3750.1
MetaRAGBase LLM=Qwen2.5-7B2026.0163.349.6
DualRAG-FTBase LLM=Qwen2.5-7B2026.0162.647.1
KiRAGBase LLM=Qwen2.5-7B2026.0162.149
DeepNoteBackbone=Qwen2.5-7B-Instruct2026.0159.9748.6
DeepNoteBackbone=Llama3.1-8B-Instruct2026.0159.5147.8
Iter-RetGenBackbone=Qwen2.5-7B-Instruct2026.0158.8147.4
DualRAGBase LLM=Qwen2.5-7B2026.0158.744.8
IRCOTBase LLM=Qwen2.5-7B2026.0156.842.4
FLAREBase LLM=Qwen2.5-7B2026.0156.441.9
Iter-RetGenBackbone=Llama3.1-8B-Instruct2026.0155.9743.8
HippoRAG2026.0255.8749.5
Search-R1Backbone=Qwen2.5-7B-Instruct2026.0154.7843.4
Vanilla RAGBackbone=Qwen2.5-14B-Instruct2026.0154.7343.8
Vanilla RAGBackbone=Qwen3-32B2026.0154.3242
QuCo-RAGBackbone=OLMo-2-32B2025.1254.241.6
Vanilla RAGBackbone=Llama3.1-8B-Instruct2026.0152.9442.2
LiR3AGBackbone=8B, Reasoning=no-think2025.1252.140.2
Vanilla RAGBackbone=Qwen2.5-7B-Instruct2026.0150.9340.4
IRCoTBackbone=Llama3.1-8B-Instruct2026.0150.4139.6
NativeRAGBase LLM=Qwen2.5-7B2026.0150.234.5
Vanilla RAGBackbone=32B, Reasoning=think2025.1250.138
IRCoTBackbone=Qwen2.5-7B-Instruct2026.0149.5536.6
QuCo-RAGBackbone=OLMo-2-13B2025.1246.835
QuCo-RAGBackbone=OLMo-2-7B2025.1246.135.3
Vanilla RAGBackbone=14B, Reasoning=think2025.1245.735.6
Vanilla RAGBackbone=32B, Reasoning=no-think2025.1245.234.6
Vanilla RAGBackbone=8B, Reasoning=think2025.1244.532.8
Vanilla RAGBackbone=14B, Reasoning=no-think2025.1244.134.2
HyperRetriever2026.0243.6542.5
REPAIR-7BAnswer generator=Llama-3.1-8B-Instruct, Context size=top-10, Retrieval protocol=rerank top-100 Contriever2026.014332.2
HyperGraphRAG2026.0242.6951
ETCBackbone=OLMo-2-32B2025.1242.230.8
REARANK-7BAnswer generator=Llama-3.1-8B-Instruct, Context size=top-10, Retrieval protocol=rerank top-100 Contriever2026.014230.2
RAPTOR2026.0241.5635.5
HyperMemorybeam width=3, depth=32026.0241.5135.5
SR-RAGBackbone=OLMo-2-7B2025.1240.729.7
SR-RAGBackbone=OLMo-2-32B2025.1240.429.5
SeaKRBackbone=OLMo-2-32B2025.1240.428.7
Vanilla RAGBackbone=8B, Reasoning=no-think2025.1240.331
FLAREBackbone=OLMo-2-32B2025.1239.828.3
SR-RAGBackbone=OLMo-2-13B2025.1239.529.7
ETCBackbone=OLMo-2-13B2025.1239.529.3
SUNAR_RAnswer generator=Llama-3.1-8B-Instruct, Context size=top-10, Retrieval protocol=rerank top-100 Contriever2026.0139.428.2
RankZephyr-7BAnswer generator=Llama-3.1-8B-Instruct, Context size=top-10, Retrieval protocol=rerank top-100 Contriever2026.0138.828.2
QwenRerank-7BAnswer generator=Llama-3.1-8B-Instruct, Context size=top-10, Retrieval protocol=rerank top-100 Contriever2026.0137.927.8
ContrieverAnswer generator=Llama-3.1-8B-Instruct, Context size=top-10, Retrieval protocol=initial retrieval2026.0137.426.4
SeaKRBackbone=OLMo-2-13B2025.1237.326.2
Rank-R1-7BAnswer generator=Llama-3.1-8B-Instruct, Context size=top-10, Retrieval protocol=rerank top-100 Contriever2026.0137.127
Direct OutputBackbone=32B, Reasoning=think2025.1235.825.4
SeaKRBackbone=OLMo-2-7B2025.123524.8
ETCBackbone=OLMo-2-7B2025.1234.725.1
Wo-RAGBackbone=OLMo-2-13B2025.1233.624.4
Wo-RAGBackbone=OLMo-2-7B2025.1231.622.6
Direct OutputBackbone=14B, Reasoning=think2025.1231.622.8
Wo-RAGBackbone=OLMo-2-32B2025.1231.322
Direct OutputBackbone=8B, Reasoning=think2025.1228.518.5
FLAREBackbone=OLMo-2-7B2025.1228.420.3
DRAGINBackbone=OLMo-2-13B2025.1227.619.5
Direct OutputBackbone=8B, Reasoning=no-think2025.1225.717.2
DRAGINBackbone=OLMo-2-7B2025.1224.717.5
Direct OutputBackbone=32B, Reasoning=no-think2025.1224.617.6
DRAGINBackbone=OLMo-2-32B2025.1224.317.7
Direct OutputBackbone=14B, Reasoning=no-think2025.1223.615
FS-RAGBackbone=OLMo-2-13B2025.1221.914.6
FLAREBackbone=OLMo-2-13B2025.1221.915.3
FS-RAGBackbone=OLMo-2-7B2025.1220.714.5
FS-RAGBackbone=OLMo-2-32B2025.1219.513.9
ToG2026.021110.08