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

Web Navigation Question Answering on WebWalkerQA

58.8Success Rate (Easy)

WebThinker-32B-RL

Updated 21d ago

Evaluation Results

Method	Links
WebThinker-32B-RL 2026.05		58.8	44.6	40.4	46.5
Qwen3-8B + ACTGUIDE-RL 2026.05		50	46.79	44.58	46.77
Qwen3-4B-Instruct + ACTGUIDE-RL 2026.05		43.75	41.78	35	39.85
Qwen3-8B + RL 2026.05		43.75	44.64	39.16	42.5
Qwen3-8B 2026.05		41.87	31.78	26.25	32.2
ARPO-8B 2026.05		26.7	33.3	29.6	30.5
Qwen2.5-7B-Instruct + ACTGUIDE-RL 2026.05		24.37	21.07	21.66	22.05
Qwen2.5-3B-Instruct + ACTGUIDE-RL 2026.05		18.75	16.07	22.08	18.82
Qwen2.5-7B-Instruct 2026.05		18.75	19.28	16.25	18.09
Qwen2.5-3B-Instruct + RL 2026.05		15	15	15.83	15.29
Qwen2.5-7B-Instruct + RL 2026.05		14.37	20.35	19.58	18.67
Qwen3-4B-Instruct + RL 2026.05		13.12	13.92	9.17	12.06
Qwen3-4B-Instruct 2026.05		8.75	3.57	0.83	3.82
Qwen2.5-3B-Instruct 2026.05		5	7.14	4.58	5.73