List your company

Spatial Reasoning on SpaCE10

91.3ScoreHuman Level
22.34840.24958.1576.051May 12, 2026

Evaluation results

MethodLinks
Human LevelModel Category=Baseline2026.0591.3
GPT-4oModel Category=Proprietary Models2026.0558.3
Claude-3.7-SonnetModel Category=Proprietary Models2026.0546
LLaVA-OneVision-7BModel Category=Open-Source General Models2026.0545.2
InternVL3-2BModel Category=Open-Source General Models2026.0544.2
SpatialForgeQwen3-VL-2BModel Category=Ours2026.0538
Qwen2.5-VL-7BModel Category=Open-Source General Models2026.0537.4
Qwen3-VL-2BModel Category=Open-Source General Models2026.0535.6
Gemini-2.0-Flash-ThinkingModel Category=Proprietary Models2026.0534.3
SpaceR-7BModel Category=Open-sourced Specialized Models2026.0533.3
SpaceQwen2.5-VL-3B-InstructModel Category=Open-sourced Specialized Models2026.0532
Qwen2.5-VL-3BModel Category=Open-Source General Models2026.0531.7
SpatialLadder-3BModel Category=Open-sourced Specialized Models2026.0527.1
SpaceMantis-8BModel Category=Open-sourced Specialized Models2026.0526.3
RandomModel Category=Baseline2026.0525