Our new X account is live! Follow @wizwand_team for updates
Home
/
Benchmarks
Puzzle Solving on Sokoban (test)
Loading...
3.74
Average Reward
turn-PPO
1.8576
2.3463
2.835
3.3237
Dec 18, 2025
Average Reward
Updated 4d ago
Evaluation Results
Method
Method
Links
Average Reward
turn-PPO
Backbone=Qwen2.5-7B
2025.12
3.74
Token-PPO
Backbone=Qwen2.5-7B
2025.12
2.9
turn-PPO
Backbone=Qwen2.5-3B
2025.12
2.29
Token-PPO
Backbone=Qwen2.5-3B
2025.12
1.93
Feedback
Search any
task
Search any
task