Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Reasoning on GPQA Diamond (P@1, P@32, #Tok)
Loading...
28.41
Precision@1
TARPO
26.7668
27.1934
27.62
28.0466
Jun 4, 2026
Precision@1
Precision@32
Token Count
Updated 1mo ago
Evaluation Results
Method
Method
Links
Precision@1
Precision@32
Token Count
TARPO
Backbone=Qwen2.5-3B-In...
2026.06
28.41
90.91
568.3
Entropy-Routed
Backbone=Qwen2.5-3B-In...
2026.06
28.3
90.91
373.2
GRPO
Backbone=Qwen2.5-3B-In...
2026.06
28.03
83.33
608.5
HRPO
Backbone=Qwen2.5-3B-In...
2026.06
27.51
88.38
603.3
Pure Latent
Backbone=Qwen2.5-3B-In...
2026.06
27.26
37.37
405.6
CoT
Backbone=Qwen2.5-3B-In...
2026.06
26.83
92.93
410.6
Feedback
Search any
task
Search any
task