List your company

Spatial decision-making on Maze size 16

64.85Success RateSpecFlow
19.318831.139442.9654.7806Jun 1, 2026

Evaluation results

MethodLinks
SpecFlowSetting=Flow(image) + AR(Text)2026.0664.85
Gemini-3-FlashSetting=Naive2026.0661.27
DiffThinkerSetting=Flow(image) Only2026.0659.42
GPT-5.1Setting=Naive2026.0658.91
MVoTSetting=Autoregressive2026.0657.81
Qwen3-VL-32BSetting=SFT2026.0646.35
Qwen3-VL-32BSetting=GRPO2026.0642.17
Qwen3-VL-8BSetting=SFT2026.0636.85
Qwen3-VL-8BSetting=GRPO2026.0633.27
Qwen3-VL-32BSetting=Naive2026.0629.84
Qwen3-VL-8BSetting=Naive2026.0621.07