List your company

Spatial Decision-Making on Maze size 8

92.71Success RateGemini-3-Flash
40.751654.240867.7381.2192Jun 1, 2026

Evaluation results

MethodLinks
Gemini-3-FlashSetting=Naive2026.0692.71
SpecFlowSetting=Flow(image) + AR(Text)2026.0692.66
SpecFlowSetting=Flow(image) + AR(Text)2026.0692.04
Gemini-3-FlashSetting=Naive2026.0691.12
DiffThinkerSetting=Flow(image) Only2026.0690.82
DiffThinkerSetting=Flow(image) Only2026.0689.95
GPT-5.1Setting=Naive2026.0689.82
MVoTSetting=Autoregressive2026.0689.14
GPT-5.1Setting=Naive2026.0688.94
MVoTSetting=Autoregressive2026.0688.61
Qwen3-VL-32BSetting=SFT2026.0687.43
Qwen3-VL-32BSetting=GRPO2026.0684.18
Qwen3-VL-8BSetting=SFT2026.0680.34
Qwen3-VL-8BSetting=GRPO2026.0676.85
Qwen3-VL-32BSetting=Naive2026.0674.05
Qwen3-VL-32BSetting=SFT2026.0670.84
Qwen3-VL-32BSetting=GRPO2026.0667.31
Qwen3-VL-8BSetting=Naive2026.0663.74
Qwen3-VL-8BSetting=SFT2026.0661.47
Qwen3-VL-8BSetting=GRPO2026.0657.88
Qwen3-VL-32BSetting=Naive2026.0654.36
Qwen3-VL-8BSetting=Naive2026.0642.75