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

Scientific Discovery on ScienceAgentBench

0.225Score

ReAct

Updated 1mo ago

Evaluation Results

Method	Links
ReAct 2026.06		0.225
LATM 2026.06		0.225
ReAct 2026.06		0.206
LATM 2026.06		0.186
ML-Master2 2026.06		0.176
ML-Master2 2026.06		0.157
ReAct 2026.06		0.147
LAMBDA 2026.06		0.127
LAMBDA 2026.06		0.118
EvoDS-evo-8B 2026.06		0.118
LAMBDA 2026.06		0.108
LATM 2026.06		0.108
AutoGen 2026.06		0.108
EvoDS-8B 2026.06		0.108
AutoGen 2026.06		0.098
Data Interpreter 2026.06		0.098
Data Interpreter 2026.06		0.088
AutoGen 2026.06		0.088
Code Interpreter 2026.06		0.088
Data Interpreter 2026.06		0.088
ML-Master2 2026.06		0.088
DataMind-7B 2026.06		0.01
DataMind-14B 2026.06		0.01
DeepAnalyze-8B 2026.06		0