Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Spatial decision-making on FrozenLake size 4
Loading...
87.94
Success Rate
SpecFlow
53.9008
62.7379
71.575
80.4121
Jun 1, 2026
Success Rate
Updated 1mo ago
Evaluation Results
Method
Method
Links
Success Rate
SpecFlow
Setting=Flow(image) +...
2026.06
87.94
Gemini-3-Flash
Setting=Naive
2026.06
86.58
DiffThinker
Setting=Flow(image) Only
2026.06
84.96
MVoT
Setting=Autoregressive
2026.06
83.31
GPT-5.1
Setting=Naive
2026.06
83.27
Qwen3-VL-32B
Setting=SFT
2026.06
79.58
Qwen3-VL-32B
Setting=GRPO
2026.06
76.43
Qwen3-VL-8B
Setting=SFT
2026.06
71.46
Qwen3-VL-8B
Setting=GRPO
2026.06
68.02
Qwen3-VL-32B
Setting=Naive
2026.06
65.47
Qwen3-VL-8B
Setting=Naive
2026.06
55.21
Feedback
Search any
task
Search any
task