List your company

Spatial decision-making on MiniBehavior size 16

79.84Success RateSpecFlow
32.592844.858957.12569.3911Jun 1, 2026

Evaluation results

MethodLinks
SpecFlowSetting=Flow(image) + AR(Text)2026.0679.84
Gemini-3-FlashSetting=Naive2026.0678.02
DiffThinkerSetting=Flow(image) Only2026.0675.21
GPT-5.1Setting=Naive2026.0673.96
MVoTSetting=Autoregressive2026.0673.88
Qwen3-VL-32BSetting=SFT2026.0660.41
Qwen3-VL-32BSetting=GRPO2026.0656.78
Qwen3-VL-8BSetting=SFT2026.0650.27
Qwen3-VL-8BSetting=GRPO2026.0646.38
Qwen3-VL-32BSetting=Naive2026.0644.28
Qwen3-VL-8BSetting=Naive2026.0634.41