List your company

Multi-hop QA Retrieval on 2WikiMultihopQA (test)

97.2R@5CoopRAG
55.91266.63177.3588.069Dec 11, 2025Dec 29, 2025Jan 17, 2026Feb 5, 2026Feb 24, 2026Mar 15, 2026Apr 3, 2026

Evaluation results

MethodLinks
CoopRAGLLM=Gemma2-27B, Single-step retrieval=true2025.1297.280.8——————
CoopRAGBackbone LLM=Gemma2-27B2025.1297.280.8——————
CoopRAGLLM=Gemma2-9B, Single-step retrieval=true2025.1296.780.1——————
CoopRAGBackbone LLM=Gemma2-9B2025.1296.780.1——————
CoopRAGLLM=GPT-4o-mini, Single-step retrieval=true2025.1296.680.4——————
CoopRAGBackbone LLM=GPT-4o-mini2025.1296.680.4——————
CoopRAGLLM=Llama3.3-70B, Single-step retrieval=true2025.1296.380.6——————
CoopRAGBackbone LLM=Llama3.3-70B2025.1296.380.6——————
BRIDGERAGCondition=C (+bridge cond.)2026.0495.27—————1.2—
BRIDGERAGCondition=B (+2-way judge)2026.0494.4—————0.3—
PROPRAG2026.0494.1———————
HippoRAG2LLM=Llama3.3-70B, Single-step retrieval=true2025.1290.476.2——————
HippoRAGBackbone LLM=Llama3.3-70B2025.1290.471.9——————
HippoRAG2Backbone LLM=Llama3.3-70B2025.1290.476.2——————
HIPPORAG22026.0490.4—————3.7—
HippoRAG2LLM=GPT-4o-mini, Single-step retrieval=true2025.1290.274.6——————
HippoRAG2Backbone LLM=GPT-4o-mini2025.1290.274.6——————
BRIDGERAGCondition=A (SVO only)2026.0489.1—————5—
HippoRAGBackbone LLM=GPT-4o-mini2025.128768.4——————
NV-Embed-v2Parameters=7B2025.1276.567.1——————
GritLM-7B2025.127667.3——————
GTE-Qwen2-7B-Instruct2025.1274.866.7——————
HopRAGLLM=GPT-4o-mini, Single-step retrieval=true2025.1270.161.7——————
HopRAGBackbone LLM=GPT-4o-mini2025.1270.161.7——————
ColBERTv22025.1268.259.2——————
GTRBackbone=T5-base2025.1267.960.2——————
SiReRAGLLM=GPT-4o-mini, Single-step retrieval=true2025.1267.660.6——————
SiReRAGBackbone LLM=GPT-4o-mini2025.1267.660.6——————
RAPTORBackbone LLM=Llama3.3-70B2025.1266.258.3——————
RAPTORBackbone LLM=GPT-4o-mini2025.126658.4——————
BM252025.1265.355.3——————
PropositionBase Retriever=ColBERTv22025.1264.955.9——————
Contriever2025.1257.546.6——————
Auto-RAGModel Parameters=7B2026.05———————66.05
BGE-largeModel Parameters=326M2026.05———————58.43
BM252026.05———————49.8
Contriever2026.05———————51.25
GritHopperModel Parameters=7B2026.05———————59.97
IRCoTModel Parameters=70B2026.05———————65.12
ITER-RETGENModel Parameters=70B2026.05———————59.96
MDRModel Parameters=110M2026.05———————53.91
R3-RAGModel Parameters=8B2026.05———————82.82
RT-RAGModel Parameters=8B2026.05———————65.14
SEALModel Parameters=406M2026.05———————48.32
SelfaskModel Parameters=70B2026.05———————47.07
Thermovariant=tune winner, α=0.4, dk=202026.03——43.526110.02——
Thermovariant=best recall, α=0.4, dk=502026.03——44.32470.003——
Thermovariant=safest, α=0.5, dk=502026.03——43.11510.001——
ThinkGRModel Parameters=8B2026.05———————93.19
True RRF†2026.03——41.21560.078——
Vector only2026.03——37.8—————
w/o RLModel Parameters=8B2026.05———————92.03
w/o SFTModel Parameters=8B2026.05———————57.7
w/o ThoughtModel Parameters=8B2026.05———————93.01