Multimodal deep search and reasoning on BrowseComp V3 benchmark leaderboard · SOTA2 Research

Multimodal deep search and reasoning on BrowseComp V3

68.03Success Rate (SR) - AvgBrowser

Evaluation Results

Method	Links
Browser2026.02		68.03	72	70	73.33	68	54	82.93	87.54	85.25	84.19	82.73	74.32
GPT-5.2-Thinking2026.02		39.13	26	48	38.67	37.33	46	66.05	61.11	79.74	54.87	71.41	64.73
OmniSeeker2026.02		36	50	28	33.33	33.33	38	57.7	67.23	55.4	49.34	60.49	58.81
OmniSeeker2026.02		33.67	42	28	37.33	38.67	18	58.44	52.94	69.7	57.46	66.27	42.42
OmniSeeker2026.02		26	22	24	25.33	34.67	20	44.66	43.7	52.11	40.73	48.72	37.94
OmniSeeker2026.02		23.67	32	24	18.67	25.33	20	47.37	50.84	68.35	40.15	43.68	39.27
Gemini-3-Pro-Preview2026.02		22.9	18	16	21.33	24	34	62.43	62.02	73.78	48.7	66.33	62.5
OmniSeeker2026.02		22.67	32	20	21.33	24	16	54.17	60.94	64.25	45.29	56.73	46.78
Claude-Sonnet-4.5-Thinking2026.02		18.33	22	16	17.33	21.33	14	47.73	58.61	58.41	34.79	54.33	35.68
OmniSeeker2026.02		16.67	18	12	10.67	29.33	10	31.33	34.84	35.04	24.45	42.43	17.76
OmniSeeker2026.02		14.33	16	14	14.67	17.33	8	26.68	28.56	35.93	22	28.36	20.05
Gemini-3-Flash-Preview2026.02		12	8	18	12	10.67	12	40.76	38.98	61.94	31.89	39.72	36.6
OmniSeeker2026.02		11.41	14	14	10.67	14.67	2	24.15	22.32	43.36	17.35	25.93	13.04
Doubao-Seed-1.82026.02		9	8	16	1.33	13.33	8	34.74	36.26	51	22.28	39.87	27.96
o4-mini2026.02		7.33	0	16	2.67	6.67	14	29.08	29.48	46.12	17.91	29.04	28.44
GPT-5.22026.02		6	0	14	4	5.33	8	25.02	17.5	44.5	19.22	25.25	21.43
OmniSeeker2026.02		5.33	2	2	9.33	8	2	13.4	8.02	19.3	15.57	15.55	6.38
Claude-Sonnet-4.52026.02		4	4	6	2.67	5.33	2	25.74	33.06	47.56	15.97	24.53	13.06
Qwen3-VL-235B-A22B-Instruct2026.02		3.33	4	4	4	2.67	2	20.52	26.34	35.38	11.69	19.64	14.39
MiMo-V2-Flash2026.02		3	2	4	2.67	4	2	8.12	4.52	17.28	5.43	9.4	4.7
GPT-4o2026.02		2.67	0	10	2.67	1.33	0	11.26	6.78	26.52	9.45	7.6	8.66
Qwen3-VL-8B-Instruct2026.02		1	0	2	1.33	0	2	6.64	3.28	15.36	6.39	4.09	5.48