List your company

General AI Assistant Tasks on GAIA (Task success rate)

71.5Task Success RatePIVOT
29.940.751.562.3May 11, 2026

Evaluation results

MethodLinks
PIVOTBackbone=Opus 4.6, Feedback regime=w/ HITL2026.0571.5
Self-CritiqueBackbone=Opus 4.62026.0568.5
PIVOTBackbone=Opus 4.5, Feedback regime=w/ HITL2026.0568.5
AgentDebugBackbone=Opus 4.62026.0566.7
PIVOTBackbone=Opus 4.6, Feedback regime=w/o HITL2026.0564.8
Self-CritiqueBackbone=Opus 4.52026.0564.2
AgentDebugBackbone=Opus 4.52026.0563.6
PIVOTBackbone=Opus 4.5, Feedback regime=w/o HITL2026.0560.6
ReACTBackbone=Opus 4.52026.0559.4
SE-AgentBackbone=Opus 4.52026.0557.6
ReACTBackbone=Opus 4.62026.0557
SE-AgentBackbone=Opus 4.62026.0556.4
PIVOTBackbone=Sonnet 4.5, Feedback regime=w/ HITL2026.0552.7
Self-CritiqueBackbone=Sonnet 4.52026.0551.5
AgentDebugBackbone=Sonnet 4.52026.0547.9
Self-CritiqueBackbone=Haiku 4.52026.0544.8
PIVOTBackbone=Haiku 4.5, Feedback regime=w/ HITL2026.0544.8
AgentDebugBackbone=Haiku 4.52026.0541.8
PIVOTBackbone=Sonnet 4.5, Feedback regime=w/o HITL2026.0541.8
PIVOTBackbone=Q3 235B, Feedback regime=w/ HITL2026.0541.8
ReACTBackbone=Sonnet 4.52026.0540.6
PIVOTBackbone=Haiku 4.5, Feedback regime=w/o HITL2026.0540.6
SE-AgentBackbone=Sonnet 4.52026.0540
Self-CritiqueBackbone=Q3 235B2026.0537.6
SE-AgentBackbone=Q3 235B2026.0537.5
ReACTBackbone=Haiku 4.52026.0537
SE-AgentBackbone=Haiku 4.52026.0537
PIVOTBackbone=Q3 235B, Feedback regime=w/o HITL2026.0535.2
AgentDebugBackbone=Q3 235B2026.0533.4
ReACTBackbone=Q3 235B2026.0531.5