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

Research Task on LiveResearchBench (LRB)

96.2Accuracy

GPT 5.4

Updated 15d ago

Evaluation Results

Method	Links
GPT 5.4 2026.05		96.2	-
Gemini 3.1 Pro 2026.05		92.5	-
LLM-guided spec search 2026.05		91	13.5
Qwen3.5-122B 2026.05		90.5	-
Best Local 2026.05		90.5	-
Claude Opus 4.6 2026.05		90	-
Qwen3.5-35B 2026.05		86.9	-
Gemma4-26B 2026.05		84.2	-
Qwen3.5-27B 2026.05		82.5	-
Nemotron-Super-120B 2026.05		80.1	-
LLM-guided spec search 2026.05		80	66.2
Qwen3.5-9B 2026.05		77.5	-
Qwen3.5-9B 2026.05		77.5	-
LLM-guided spec search 2026.05		75	62.5
LLM-guided spec search 2026.05		68.5	17.5
Granite 3.3 8B 2026.05		60.3	-
Gemma4-E4B 2026.05		51	-
Gemma4-E4B 2026.05		51	-
Granite 4.0 H-Small 2026.05		50.6	-
Qwen3.5-4B 2026.05		13.8	-
Nemotron-Nano-4B 2026.05		12.5	-