Share your thoughts, 1 month free Claude Pro on usSee more

Spatial decision-making on Maze, MiniBehavior, and FrozenLake Average

79.15Success Rate (%)

SpecFlow

Updated 1mo ago

Evaluation Results

Method	Links
SpecFlow 2026.06		79.15
Gemini-3-Flash 2026.06		77.67
DiffThinker 2026.06		75.59
GPT-5.1 2026.06		74.53
MVoT 2026.06		74.1
Qwen3-VL-32B 2026.06		63.27
Qwen3-VL-32B 2026.06		59.83
Qwen3-VL-8B 2026.06		54.25
Qwen3-VL-8B 2026.06		50.73
Qwen3-VL-32B 2026.06		48.02
Qwen3-VL-8B 2026.06		38.11