Share your thoughts, 1 month free Claude Pro on usSee more

Question Answering on NQ (Average Performance Score)

85.7Average Performance Score

DeepSeek-R1 w/ GuarantRAG

Updated 8d ago

Evaluation Results

Method	Links
DeepSeek-R1 w/ GuarantRAG 2026.04		85.7
DeepSeek-V3 w/ GuarantRAG 2026.04		84.1
DeepSeek-R1 w/ P-RAG 2026.04		84.1
Gemini-2.5-Pro w/ GuarantRAG 2026.04		83.4
DeepSeek-V3 w/ P-RAG 2026.04		82.7
DeepSeek-R1 w/ SelfRAG 2026.04		82.6
Claude-3.7-Sonnet w/ GuarantRAG 2026.04		82.3
Gemini-2.5-Pro w/ P-RAG 2026.04		81.8
DeepSeek-V3 w/ SelfRAG 2026.04		81.4
DeepSeek-R1 w/ Standard RAG 2026.04		81.2
Claude-3.7-Sonnet w/ P-RAG 2026.04		80.7
Gemini-2.5-Pro w/ SelfRAG 2026.04		80.3
DeepSeek-V3 w/ Standard RAG 2026.04		79.8
Claude-3.7-Sonnet w/ SelfRAG 2026.04		79.1
DeepSeek-R1 2026.04		78.9
Gemini-2.5-Pro w/ Standard RAG 2026.04		78.9
Claude-3.7-Sonnet w/ Standard RAG 2026.04		77.6
DeepSeek-V3 2026.04		77.2
Gemini-2.5-Pro 2026.04		76.4
Claude-3.7-Sonnet 2026.04		75.1