Vision-Language Task Evaluation on DevCV Toolbox Ego4D out-of-domain benchmark leaderboard · SOTA2 Research

Vision-Language Task Evaluation on DevCV Toolbox Ego4D out-of-domain

93.5Overall AccuracyHuman performance

Evaluation Results

Method	Links
Human performance2025.12		93.5	96.4	98.2	96.4	96.4	98.8	90.9	100	58.2	100	100	92.7
Gemini-2.5-pro2025.12		88.2	81.9	88	94.8	91.9	100	90.2	91.3	50.3	87.9	96.5	97.8
GPT-52025.12		86.7	77.5	88	96.8	91.9	100	88.7	94.4	50.3	82.6	94.6	88.5
Gemini-2.5-flash2025.12		77.7	72.9	49.6	86.7	92.5	99.2	88.4	80.6	37.1	70.2	97.8	80.1
GPT-4o2025.12		67.6	62.1	45.1	94.7	85.3	100	80.4	45.5	13.2	48.3	84.3	84.6
Qwen2.5-VL-3B2025.12		48.1	35.7	32.6	44.1	41.9	25.7	86.7	79.8	28.9	51.1	50.2	53.4
InternVL3.5-1B2025.12		43.7	34.7	34	34.1	33.8	24.9	60.7	73.9	16.9	68.5	49	49.9
BabyLLaVA-V22025.12		41.1	33.9	32.9	42.4	29.8	40.7	30	55.3	17.7	37.1	86	45.8
LLaVA-OneVision-0.5B2025.12		39.4	43.9	32.6	33.3	27.7	22.6	21.6	73	15.2	67.7	46.8	49.4
Random guess2025.12		31.8	8.33	33.3	33.3	25	25	25	50	12.5	37.5	50	50