Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Arithmetic Reasoning on Countdown (val)
Loading...
76.27
Validation Reward (mean@4)
ACPO
4.1252
22.8551
41.585
60.3149
Jun 21, 2026
Validation Reward (mean@4)
Updated 1mo ago
Evaluation Results
Method
Method
Links
Validation Reward (mean@4)
ACPO
Backbone=Qwen2.5-3B-In...
2026.06
76.27
ACPO
Backbone=Qwen2.5-3B-In...
2026.06
75.74
DAPO
Backbone=Qwen2.5-3B-In...
2026.06
74.38
CISPO
Backbone=Qwen2.5-3B-In...
2026.06
74.25
Low-Entropy
Backbone=Qwen2.5-3B-In...
2026.06
74.04
High-Entropy
Backbone=Qwen2.5-3B-In...
2026.06
73.76
DAPO
Backbone=Qwen2.5-3B-In...
2026.06
73.27
High-Entropy
Backbone=Qwen2.5-3B-In...
2026.06
70.58
AR-Lopti
Backbone=Qwen2.5-3B-In...
2026.06
65.82
Low-Entropy
Backbone=Qwen2.5-3B-In...
2026.06
61.81
AR-Lopti
Backbone=Qwen2.5-3B-In...
2026.06
57.41
CISPO
Backbone=Qwen2.5-3B-In...
2026.06
55.12
Base Model
Backbone=Qwen2.5-3B-In...
2026.06
6.9
Feedback
Search any
task
Search any
task