Spatial Reasoning on SPAR-Bench tiny benchmark leaderboard · SOTA2 Research

Spatial Reasoning on SPAR-Bench tiny

72.32Medium Difficulty ScoreHuman Level

Evaluation Results

Method	Links
Human Level2025.12		72.32	67.27	55.31	72.75	74.25	28.75	36.25	78.25	52.25	66.5	33.5	92	64	60.97	76.22	80	94	70	92	80	78	82	50	60
Human Level2026.03		72.32	67.27	55.31	72.75	74.25	28.75	36.25	78.25	52.25	66.5	33.5	92	64	60.97	76.22	80	94	70	92	80	78	82	50	60
Chance Level (Frequency)2025.12		38.25	32.74	31.19	43.09	43.51	17.38	13.05	41.9	30.99	27.4	32.17	29.01	26.75	59	32.29	52.94	50.6	28.25	26.92	26.59	26.34	26.74	26.49	25.77
GPT-4o2025.12		24.93	36.39	29.25	53.8	45	15	13.6	37.4	34.4	23.4	24.4	30	16	28.8	45.11	64	64	58	46	46	32	44	30	22
GPT-4o2026.03		24.93	36.39	29.25	53.8	45	15	13.6	37.4	34.4	23.4	24.4	30	16	28.8	45.11	64	64	58	46	46	32	44	30	22
Qwen2-VL-72B2025.12		23.39	35.62	35.28	45.4	49.8	13.8	10	54.6	49.4	36.8	22.4	42	18	10.16	40	60	68	50	38	44	18	28	18	36
Qwen2-VL-72B2026.03		23.39	35.62	35.28	45.4	49.8	13.8	10	54.6	49.4	36.8	22.4	42	18	10.16	40	60	68	50	38	44	18	28	18	36
Qwen2.5-VL-72B2025.12		23.05	39.4	35.35	53.2	46.8	17.8	29	49.6	57.4	14.4	14.6	40	16	13.16	48.44	74	74	60	56	50	20	34	24	44
Qwen2.5-VL-72B2026.03		23.05	39.4	35.35	53.2	46.8	17.8	29	49.6	57.4	14.4	14.6	40	16	13.16	48.44	74	74	60	56	50	20	34	24	44
Chance Level (Random)2025.12		22.65	—	—	—	—	—	—	—	—	—	—	24.5	—	25.09	23.82	22.02	31.25	25.27	22.16	25.81	24.42	24.17	26.89	—
Claude-3.7-Sonnet2025.12		7.33	21.77	25.43	41	45.4	11.2	12.2	42.6	19.6	26	5.4	16	6	0	23.33	40	48	22	36	14	12	20	6	12
Claude-3.7-Sonnet2026.03		7.33	21.77	25.43	41	45.4	11.2	12.2	42.6	19.6	26	5.4	16	6	0	23.33	40	48	22	36	14	12	20	6	12