List your company

Tool-use Agent on GAIA

64.1Text Success RateAutoAgent
8.56422.98237.451.818Mar 10, 2026

Evaluation results

MethodLinks
AutoAgentBackbone=gemini-3-pro2026.0364.15031.654.5
AutoAgentBackbone=gemini-3-pro-thinking2026.0362.158.323.752.7
Direct AnswerBackbone=gemini-3-pro-thinking2026.0344.72513.234.5
AutoAgentBackbone=gpt-4o2026.0343.72531.638.2
Direct AnswerBackbone=gemini-3-pro2026.0342.716.715.832.7
DeepAgentBackbone=gpt-4o2026.0342.524.631.637.6
DeepAgentBackbone=gemini-3-pro2026.0338.816.72.627.3
ReActBackbone=gemini-3-pro2026.0336.916.734.233.3
ReActBackbone=gemini-3-pro-thinking2026.033212.539.530.9
DeepAgentBackbone=gemini-3-pro-thinking2026.033220.8023
ReActBackbone=gpt-4o2026.0326.220.826.325.5
Direct AnswerBackbone=gpt-4o2026.0310.702.67.3