Short-horizon Ego-Centric Spatiotemporal Reasoning on Dual-Level Evaluation Framework Short level benchmark leaderboard · SOTA2 Research

Short-horizon Ego-Centric Spatiotemporal Reasoning on Dual-Level Evaluation Framework Short level

97.3Acc (Frame Interval 5)HUMAN PERCEPTION

Evaluation Results

Method	Links
HUMAN PERCEPTION2026.04		97.3	97.3	98.2	99.1	99.1	100	100	100	98.9
EGOTSR-TAG2026.04		82.4	84.8	86.2	87.1	87.7	88.6	88.6	88.7	86.8
EGOTSR-LONGTAG2026.04		80.8	86.3	89.7	88.8	89.8	87.2	89	88.2	87.5
GEMINI-2.5-PRO2026.04		71.3	75	72.8	75	77.2	80.9	89.7	100	80.2
GPT-4O2026.04		71.1	71.7	76.7	77.8	79.2	79.2	79.2	79.2	76.8
GPT-O4MINI2026.04		70	70.3	72.8	67.7	60	67.5	67.7	70.3	68.3
EGOTSR-COT2026.04		68.3	70.2	72.2	73.7	74.5	75.6	76.7	77.8	73.6
GPT-O32026.04		67.2	61.4	59.6	61.6	59.6	64	62	66.6	62.8
SEED-1.62026.04		66.2	66.8	70.6	73.3	83.2	82.9	86.8	98.3	78.5
DEEPSEEK-VL22026.04		64.4	66.9	65.2	67.8	67	69.1	77.4	81	69.9
DOUBAO-1.5-PRO2026.04		64	76	56	46	56.9	86	90	96	71.4
LLAVA-ONEVIS2026.04		54.8	55.4	53.5	49.8	52.1	50.6	52.7	50.4	52.5
QWEN2.5-VL-7B2026.04		53.1	55.4	55	57.3	55.2	56.1	56	56.5	55.6
QWEN2.5-VL-3B2026.04		51.3	54.2	57.1	58.8	54.5	59.6	59.6	66.7	57.7
VIDEO-3D LLM2026.04		50	50.2	50.7	51.1	49.4	50.4	51.1	49.3	50.3
INTERNVL-8B2026.04		47.7	50.6	49.4	47.9	47.1	48.7	48.9	51	48.9
NVILA-8B2026.04		46.5	52	46	56.1	50	50	47.5	55	50.4
PALIGEMMA2026.04		43.8	45.1	44.2	45.7	42.3	43.6	42.5	46.1	44.2
3D-LLAVA2026.04		43.2	46.1	46.4	47.9	49.6	48.8	48	47.9	47.2
LL3DA2026.04		36	33.9	32.9	34.5	36.9	33.3	39.6	36	35.4
CHAT-SCENE2026.04		30.8	32.5	40.7	49	41.6	45.3	43.3	51	41.8
3D-LLM2026.04		27	28	30.5	30.3	33.9	32.5	36.8	37.7	32.1