Share your thoughts, 1 month free Claude Pro on usSee more

Visual Spatial Intelligence on VSI-Bench

79.2Average Score

Human Level†

Updated 23d ago

Evaluation Results

Method
Human Level† 2026.03	79.2	94.3	47	60.4	45.9	94.7	95.8	95.8	100
OmniStream-7B 2026.03	70.6	73.2	55.7	76.9	74.8	72.3	82.1	45.4	84.6
SpaceMind 2026.03	69.6	73.3	61.4	77.3	74.2	67.2	88.4	44.3	70.5
Cog3DMap-8B 2026.03	65.1	69.6	54.8	67.8	67.1	64.8	85.6	43	67.9
Pixelis (Qwen3-VL-8B-Instruct) 2026.03	64.4	-	-	-	-	-	-	-	-
InternVL3.5-A3B (with tools) 2026.03	63.7	-	-	-	-	-	-	-	-
Qwen3-VL-30B-A3B-Instruct 2026.03	63.2	-	-	-	-	-	-	-	-
Pixelis: SFT + RFT 2026.03	62.8	-	-	-	-	-	-	-	-
Pixelis: SFT + TTRL 2026.03	62.8	-	-	-	-	-	-	-	-
PRM (process reward; tools; 8B) 2026.03	62.3	-	-	-	-	-	-	-	-
Pixelis: RFT + TTRL 2026.03	61.9	-	-	-	-	-	-	-	-
VST-7B 2026.03	61.2	71.6	43.8	75.5	69.2	60	55.6	44.3	69.2
VLM-3R-7B 2026.03	60.9	70.2	49.4	69.2	67.1	65.4	80.5	45.4	40.1
VLM-3R-7B 2026.03	60.9	70.2	49.4	69.2	67.1	65.4	80.5	45.4	40.1
Pixelis: RFT only 2026.03	60.8	-	-	-	-	-	-	-	-
VST-7B 2026.03	60.6	72	44.4	74.3	68.3	59.7	55.8	44.9	65.2
Pixel Reasoner (Qwen3-VL-8B-Instruct) 2026.03	60.1	-	-	-	-	-	-	-	-
Pixelis: SFT only 2026.03	60.1	-	-	-	-	-	-	-	-
Pixelis: TTRL only 2026.03	59.6	-	-	-	-	-	-	-	-
Qwen3-VL-8B-Instruct 2026.03	59.4	-	-	-	-	-	-	-	-
VG-LLM-8B 2026.03	50.7	67.9	37.7	58.6	62	46.6	40.7	32.4	59.2
Spatial-MLLM-4B 2026.03	48.4	65.3	34.8	63.1	45.1	41.3	46.2	33.5	46.3
VG-LLM-4B 2026.03	47.3	66	37.8	55.2	59.2	44.6	45.6	33.5	36.4
3DRS-7B 2026.03	45.9	68.7	34.8	53.6	56.6	40.9	43.2	30.4	39.2
Gemini-1.5-Pro 2026.03	45.4	56.2	30.9	64.1	43.6	51.3	46.3	36	34.6
VILASR-7B 2026.03	45.4	63.5	34.4	60.6	30.9	48.9	45.2	30.4	49.2
Gemini-1.5-Pro 2026.03	45.4	56.2	30.9	64.1	43.6	51.3	46.3	36	34.6
SpaceR-7B 2026.03	43.5	61.9	28.6	60.9	35.2	38.2	46	31.4	45.6
Gemini-1.5-Flash 2026.03	42.1	49.8	30.8	53.5	54.4	37.7	41	31.5	37.8
SPAR-8B 2026.03	41.1	-	-	-	-	-	-	-	-
LLaVA-NeXT-Video-72B 2026.03	40.9	48.9	22.8	57.4	35.3	42.4	36.7	35	48.6
LLaVA-OneVision-72B 2026.03	40.2	43.5	23.9	57.6	37.5	42.5	39.9	32.5	44.6
InternVL2-40B 2026.03	36	34.9	26.9	46.5	31.8	42.1	32.2	34	39.6
LLaVA-Video-7B 2026.03	35.6	48.5	14	47.8	24.2	43.5	42.4	34	30.6
LLaVA-NeXT-Video-7B 2026.03	35.6	48.5	14	47.8	24.2	43.5	42.4	34	30.6
InternVL2-8B 2026.03	34.6	23.1	28.7	48.2	39.8	36.7	30.7	29.9	39.6
GPT-4o 2026.03	34	46.2	5.3	43.8	38.2	37	41.3	31.5	28.5
GPT-4o 2026.03	34	46.2	5.3	43.8	38.2	37	41.3	31.5	28.5
Qwen2.5-VL-7B 2026.03	32.7	34.5	19.4	47.6	40.8	32.8	24.5	32.5	29.4
LLaVA-OneVision-7B 2026.03	32.4	47.7	20.2	47.4	12.3	42.5	35.2	29.4	24.4
LLaVA-OneVision-7B 2026.03	32.4	47.7	20.2	47.4	12.3	42.5	35.2	29.4	24.4
Gemini-2.5-Pro (minimal) 2026.03	31.4	-	-	-	-	-	-	-	-
VILA-1.5-40B 2026.03	31.2	22.4	24.8	48.7	22.7	40.5	25.7	31.5	32.9
LongVA-7B 2026.03	29.2	38	16.6	38.9	22.2	33.1	43.3	25.4	15.7
VILA-1.5-8B 2026.03	28.9	17.4	21.8	50.3	18.8	32.1	34.8	31	24.8
LLaVA-OneVision-0.5B 2026.03	28	46.1	28.4	15.4	28.3	28.9	36.9	34.5	5.8
InternVL2-2B 2026.03	27.4	21.8	24.9	22	35	33.8	44.2	30.5	7.1
LongVILA-8B 2026.03	21.6	29.1	9.1	16.7	0	29.6	30.7	32.5	25.5
Random 2026.03	-	-	-	-	-	25	36.1	28.3	25
SAT-LLaVA-Video-7B 2026.03	-	-	-	-	47.3	41.1	37.1	36.1	40.4