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

Interactive Decision-making on WebShop (Real, WM, W2R, CR metrics)

39Real

GPT-5

Updated 3d ago

Evaluation Results

Method	Links
GPT-5 2026.04		39	39	35.5	0.91	73
GPT-5 2026.04		39	43.7	41.5	1.064	75.6
GPT-5 2026.04		39	43.5	37.5	0.962	75.6
GPT-5 2026.04		39	36	34.5	0.88	69
GPT-5 2026.04		39	44	41	1.051	76.9
GPT-5 2026.04		39	41.5	36.5	0.94	72
GPT-4o 2026.04		19	19	17.5	0.92	76
GPT-4o 2026.04		19	22.8	18.5	0.974	76.3
GPT-4o 2026.04		19	21	20	1.05	84
GPT-4o 2026.04		19	19	17.5	0.92	71
GPT-4o 2026.04		19	23.5	22	1.158	81.6
GPT-4o 2026.04		19	21.5	21	1.11	89
Qwen3-32B 2026.04		16.5	19	14.5	0.88	45.5
Qwen3-32B 2026.04		16.5	14.4	12	0.727	42.4
Qwen3-32B 2026.04		16.5	15.5	15	0.91	48.5
Qwen3-32B 2026.04		16.5	13.5	13	0.79	48.5
Qwen3-32B 2026.04		16.5	17.5	12	0.727	36.4
Qwen3-32B 2026.04		16.5	17.5	14	0.85	51.5
Qwen3-8B 2026.04		14.5	16.5	12	0.83	34.5
Qwen3-8B 2026.04		14.5	15.2	9.5	0.655	31
Qwen3-8B 2026.04		14.5	17	13.5	0.93	48.3
Qwen3-8B 2026.04		14.5	27.5	12	0.83	34.5
Qwen3-8B 2026.04		14.5	14.9	11.5	0.793	27.6
Qwen3-8B 2026.04		14.5	12.5	10.5	0.72	34.5