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

Failure Attribution on Who&When

60.79Agent Accuracy

AEA-4B

Updated 8d ago

Evaluation Results

Method	Links
AEA-4B 2026.05		60.79	32.7
Claude 3.7 Sonnet 2026.05		58.7	24.6
Qwen2.5-72B 2026.05		55.56	10.32
Qwen3-32B 2026.05		53.97	15.87
GPT-4o 2026.05		50	30.95
Llama-3.1-70B 2026.05		48.41	16.67
AEA-4B 2026.05		48.41	25.87
Claude 3.7 Sonnet 2026.05		46.03	30.95
Qwen3-32B 2026.05		38.1	23.02
GPT-4o-mini 2026.05		35.71	19.84
GPT-4o 2026.05		33.33	26.43
Qwen3-8B 2026.05		31.75	5.56
Llama-3.1-70B 2026.05		30.95	19.84
GPT-4o-mini 2026.05		30.95	19.05
Llama-3.1-8B 2026.05		24.6	13.49
Claude 3.5 Haiku 2026.05		21.43	16.67
Qwen2.5-72B 2026.05		20.63	16.67
Qwen3-8B 2026.05		20.63	10.32
Claude 3.5 Haiku 2026.05		12.7	7.94
Qwen2.5-7B 2026.05		11.11	6.35
Qwen2.5-7B 2026.05		10.32	5.56
Llama-3.1-8B 2026.05		9.52	2.38