Our new X account is live! Follow @wizwand_team for updates

Spatial Reasoning on VSI-Bench 1.0 (test)

25Relative Distance Error

Chance Level (Random)

Updated 3d ago

Evaluation Results

Method
Chance Level (Random) 2026.01	25	-	-	-	-	-	36.1	28.3	25
Chance Level (Frequency) 2026.01	25.1	34	62.1	32	29.9	33.1	47.9	28.4	25.2
LLaVA-OneVision-0.5B 2026.01	28.9	28	46.1	28.4	15.4	28.3	36.9	34.5	5.8
LongVILA-8B 2026.01	29.6	21.6	29.1	9.1	16.7	0	30.7	32.5	25.5
VILA-1.5-8B 2026.01	32.1	28.9	17.4	21.8	50.3	18.8	34.8	31	24.8
VILA-1.5-8B 2026.02	32.1	28.9	17.4	21.8	50.3	18.8	34.8	31	24.8
LongVA-7B 2026.01	33.1	29.2	38	16.6	38.9	22.2	43.3	25.4	15.7
InternVL2-2B 2026.01	33.8	27.4	21.8	24.9	22	35	44.2	30.5	7.1
InternVL2-8B 2026.01	36.7	34.6	23.1	28.7	48.2	39.8	30.7	29.9	39.6
GPT-4o 2026.01	37	34	46.2	5.3	43.8	38.2	41.3	31.5	28.5
GPT-4o 2026.02	37	34	46.2	5.3	43.8	38.2	41.3	31.5	28.5
Gemini-1.5 Flash 2026.01	37.7	42.1	49.8	30.8	53.5	54.4	41	31.5	37.8
Gemini-1.5-Flash 2026.02	37.7	42.1	49.8	30.8	53.5	54.4	41	31.5	37.8
InternVL2-8B 2026.02	38	37.5	31.3	29	48.9	44.2	33.4	28.9	46.4
Qwen2.5VL-3B 2026.02	38.3	30.6	24.3	24.7	31.7	22.6	41.6	26.3	21.2
Qwen2.5VL-7B 2026.02	38.6	33	40.9	14.8	43.4	10.7	38.5	33	29.8
VILA-1.5-40B 2026.01	40.5	31.2	22.4	24.8	48.7	22.7	25.7	31.5	32.9
Spatial-MLLM-4B 2026.01	41.3	48.4	65.3	34.8	63.1	45.1	46.2	33.5	46.3
Spatial-MLLM-4B 2026.02	41.3	48.4	65.3	34.8	63.1	45.1	46.2	33.5	46.3
InternVL2-40B 2026.01	42.1	36	34.9	26.9	46.5	31.8	32.2	34	39.6
LLaVA-NeXT-Video-72B 2026.01	42.4	40.9	48.9	22.8	57.4	35.3	36.7	35	48.6
LLaVA-OneVision-7B 2026.01	42.5	32.4	47.7	20.2	47.4	12.3	35.2	29.4	24.4
LLaVA-OneVision-72B 2026.01	42.5	40.2	43.5	23.9	57.6	37.5	39.9	32.5	44.6
LLaVA-OneVision-7B 2026.02	42.5	32.4	47.7	20.2	47.4	12.3	35.2	29.4	24.4
LLaVA-NeXT-Video-7B 2026.01	43.5	35.6	48.5	14	47.8	24.2	42.4	34	30.6
LLaVA-Video-7B 2026.02	43.5	35.6	48.5	14	47.8	24.2	42.4	34	30.6
VG-LLM-4B 2026.02	44.6	47.3	66	37.8	55.2	59.2	45.6	33.5	36.4
VG-LLM-8B 2026.02	46.6	50.7	67.9	37.7	58.6	62	40.7	32.4	59.2
Gemini-1.5 Pro 2026.01	51.3	45.4	56.2	30.9	64.1	43.6	46.3	36	34.6
Gemini-1.5-Pro 2026.02	51.3	45.4	56.2	30.9	64.1	43.6	46.3	36	34.6
Ours (25% dataset) 2026.01	52.1	58.8	68.2	53	64.4	69.1	70.9	38.1	54.5
Ours (100% dataset) 2026.01	56.9	61	70.8	55	70.1	69.4	69.8	38.1	57.4
Spa3-VLM-4B 2026.02	57.9	58.6	69	39.3	70.6	62.8	59.5	36.1	73.6
VLM-3R (25% dataset) 2026.01	58.4	54	69.2	41.7	67.6	62.4	47.3	43.3	41.7
VG-LLM-8B 2026.01	63.8	58.2	71.7	53.8	68.8	62.1	83	44.3	18.4
Cambrian-S-3B 2026.02	64.8	57.3	70.7	40.6	68	46.3	61.9	27.3	78.8
VLM-3R (100% dataset) 2026.01	65.4	60.9	70.2	49.4	69.2	67.1	80.5	45.4	40.1