Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Mathematical Reasoning on RLVR Math Reasoning Suite (test)
Loading...
82.8
Accuracy (ACC)
GRPO + SALT
31.216
44.608
58
71.392
Jun 4, 2026
Accuracy (ACC)
Pass@8 Rate
Updated 1mo ago
Evaluation Results
Method
Method
Links
Accuracy (ACC)
Pass@8 Rate
GRPO + SALT
Model=Qwen-3-8B, Budge...
2026.06
82.8
92.3
GRPO
Model=Qwen-3-8B, Budge...
2026.06
80.1
89
GRPO + SALT
Model=Deepseek-Distill...
2026.06
69.7
88.1
GRPO
Model=Deepseek-Distill...
2026.06
66.9
85.8
GRPO + SALT
Model=Deepseek-Distill...
2026.06
53.7
78.1
GRPO
Model=Deepseek-Distill...
2026.06
51.2
75.4
GRPO + SALT
Model=Qwen-2.5-Math-1....
2026.06
46.9
65.1
GRPO
Model=Qwen-2.5-Math-1....
2026.06
43.7
62.5
GRPO + SALT
Model=Deepseek-Math-7B...
2026.06
37.2
55.7
GRPO + SALT
Model=LLaMA-3.1-8B, Bu...
2026.06
36.1
57
GRPO
Model=LLaMA-3.1-8B, Bu...
2026.06
33.5
54.8
GRPO
Model=Deepseek-Math-7B...
2026.06
33.2
52.2
Feedback
Search any
task
Search any
task