Choice-based Visual Question Answering on EgoPoint-Bench real-world 1.0 (test) benchmark leaderboard · SOTA2 Research

Choice-based Visual Question Answering on EgoPoint-Bench real-world 1.0 (test)

96.24BPHuman

Evaluation Results

Method	Links
Human2026.04		96.24	98.04	96.39	95.65	89.09	96
Gemini 3 Flash2026.04		67.04	73.98	78.89	80.9	63.02	71.84
Gemini 3 Pro2026.04		66.63	75.44	79.06	83.28	60.16	72
Qwen3-VL-32B2026.04		62.09	71.35	81.96	73.43	83.81	71.84
Qwen3-VL-8B2026.04		60.36	72.28	81.96	71.94	88.57	71.96
Qwen3-VL-32B2026.04		56.38	65.03	76.09	79.7	56.83	64.3
GPT-5.2 Instant2026.04		55.31	67.49	81.62	69.55	71.27	66.76
InternVL3.5-14B2026.04		54.03	73.1	80.26	68.66	82.86	68.92
GLM-4.6V-Flash2026.04		53.88	60.7	66.55	67.16	72.7	61.26
GPT-5 mini2026.04		52.81	66.73	67.32	66.27	52.38	60.57
InternVL3.5-8B2026.04		50.56	69.88	74.47	63.88	90	66.13
InternVL3.5-8B2026.04		50.05	60.88	63.32	68.96	50.79	57.09
GLM-4.6V-Flash2026.04		48.32	59.77	67.32	72.84	43.49	56.42
Qwen3-VL-8B2026.04		47.81	58.42	74.55	68.96	53.17	58.14
InternVL3.5-14B2026.04		47.76	65.09	72.51	65.07	45.24	58.59
InternVL3.5-2B2026.04		46.33	64.04	71.83	57.31	89.68	62.03
InternVL3.5-2B2026.04		44.85	60.47	62.55	59.4	43.65	53.73
LLaVA-NeXT-7B2026.04		40.1	66.32	71.23	40.9	90.63	59.64
LLaVA-1.5-7B2026.04		37.5	56.55	64.17	33.43	95.4	54.54
LLaVA-1.5-7B2026.04		36.48	45.85	62.13	22.69	69.37	47.19
LLaVA-NeXT-7B2026.04		31.68	51.75	60.09	39.4	56.19	46.44
Random2026.04		25.19	22.74	37.3	26.32	45.76	28.94