List your company

Spatial Reasoning on ReasonMap

55.1AccuracyGemini-2.5-Pro
-0.64413.82828.342.772Jun 11, 2026

Evaluation results

MethodLinks
Gemini-2.5-ProModel Type=Commercial2026.0655.1
GPT-5Model Type=Commercial2026.0653.8
w/o SFTModel Type=Ablation2026.0645.1
Gemini-2.5-FlashModel Type=Commercial2026.0643.2
Qwen3-235BModel Type=Open-Source2026.0630.7
PERIA-8BModel Type=Proposed2026.0627.1
w/o RLModel Type=Ablation2026.0621.1
Qwen3-32BModel Type=Open-Source2026.0620.6
Qwen3-30BModel Type=Open-Source2026.0618.2
w/o ToolModel Type=Ablation2026.0613.7
Qwen3-8BModel Type=Backbone Baseline2026.0612.3
R1-OneVision-7BModel Type=Tool-Augmented2026.068.2
InternVL3.5-8BModel Type=Open-Source2026.065.8
InternVL3.5-14BModel Type=Open-Source2026.064.4
Mini-o3Model Type=Tool-Augmented2026.061.6
VTool-R1-7BModel Type=Tool-Augmented2026.061.5