List your company

Spatial relationship understanding on MultihopSpatial

37.68AccuracyThinking with Grounding Box
3.474412.354721.23530.1153Jun 15, 2026

Evaluation results

MethodLinks
Thinking with Grounding BoxModel scale=4B, Grounding interface=Box, Grounding reward=w/2026.0637.6866.4
Thinking with Grounding PointModel scale=4B, Grounding interface=Point, Grounding reward=w/o2026.0637.0365.4
Thinking with Grounding PointModel scale=4B, Grounding interface=Point, Grounding reward=w/2026.0637.0165.02
Thinking with Grounding BoxModel scale=4B, Grounding interface=Box, Grounding reward=w/o2026.0634.8963.82
Gemma3-27B-ITModel scale=27B2026.0630.9445.82
Gemma3-12B-ITModel scale=12B2026.0630.0843.58
Gemma3-4B-ITModel scale=4B2026.0622.736.87
Non-grounded ThinkingModel scale=4B, Grounding interface=None2026.064.7911.67