List your company

Automating Agent Evaluation on AgentEvalBench

65Eval@1EvalAgent
1326.54053.5May 12, 2026

Evaluation results

MethodLinks
EvalAgentLLM=Sonnet 4.52026.0565
EvalAgentLLM=Haiku 4.52026.0562.5
Agent-Sourcecode (B2)LLM=Sonnet 4.52026.0560
Agent-Sourcecode (B2)LLM=Haiku 4.52026.0545
Agent-Onestage (B3)LLM=Sonnet 4.52026.0535
Agent-Twostage (B4)LLM=Haiku 4.52026.0532.5
Agent-Twostage (B4)LLM=Sonnet 4.52026.0530
Agent-Onestage (B3)LLM=Haiku 4.52026.0517.5
LLM-Singleturn (B1)LLM=Sonnet 4.52026.0517.5
LLM-Singleturn (B1)LLM=Haiku 4.52026.0515