List your company

Multi-hop Factual Reasoning on FRAMES

82.3AccuracyROMA
28.01242.10656.270.294Feb 2, 2026

Evaluation results

MethodLinks
ROMASearch=true, Backbone=GLM-4.6, Source_Type=Open-source2026.0282.3
Kimi-ResearcherSearch=true, Source_Type=Open-source2026.0278.8
Open Deep Search (DeepSeek-R1)Search=true, Source_Type=Open-source2026.0275.3
GLM-4.6Search=true, Source_Type=Open-source2026.0271.2
GPT-4o Search PreviewSearch=true, Source_Type=Closed-source2026.0265.6
GPT-4oSearch=false, Source_Type=Closed-source2026.0250.5
Perplexity Sonar Reasoning ProSearch=true, Source_Type=Closed-source2026.0244.4
PerplexitySearch=true, Source_Type=Closed-source2026.0242.4
Llama-3.1-70BSearch=false, Source_Type=Open-source2026.0234.3
DeepSeek-R1Search=false, Source_Type=Open-source2026.0230.1