Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Video Spatial Reasoning on Video-RoboSpatial
Loading...
92.1
Configuration Score
Human Level
46.236
58.143
70.05
81.957
Mar 5, 2026
Configuration Score
Updated 2mo ago
Evaluation Results
Method
Method
Links
Configuration Score
Human Level
Size=–
2026.03
92.1
Thinking with Spatial Code
Size=4B
2026.03
67
Thinking with Spatial Code + 2D box
Size=4B
2026.03
67
Thinking with Spatial Code (w/o RL)
Size=4B
2026.03
65.3
Thinking with Spatial Code (w/o RL) + 2D box
Size=4B
2026.03
65.3
GPT-5
Size=–
2026.03
60.3
LLaVA-Video
Size=72B
2026.03
58
SpaceR
Size=7B
2026.03
56
LLaVA-OneVision
Size=72B
2026.03
56
Seed-1.6
Size=230B
2026.03
54.3
LLaVA-OneVision
Size=7B
2026.03
54.3
Gemini-2.5-Pro
Size=–
2026.03
53.3
GPT-4o
Size=–
2026.03
53
Qwen3-VL
Size=4B
2026.03
52.7
Qwen3-VL + 2D box
Size=4B
2026.03
52.7
LLaVA-Video
Size=7B
2026.03
52
Qwen2.5-VL
Size=7B
2026.03
49.7
SpatialLadder
Size=3B
2026.03
49.3
Spatial-MLLM
Size=4B
2026.03
49
Qwen3-VL
Size=8B
2026.03
48
Feedback
Search any
task
Search any
task