Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Math Reasoning on GSM8K (Ngen Exact Match Accuracy)
Loading...
75
Exact Match Accuracy (Ngen=128)
DISPO
66.16
68.455
70.75
73.045
Feb 6, 2026
Exact Match Accuracy (Ngen=128)
Exact Match Accuracy (Ngen=256)
Exact Match Accuracy (Ngen=512)
Average Exact Match Accuracy
Updated 4mo ago
Evaluation Results
Method
Method
Links
Exact Match Accuracy (Ngen=128)
Exact Match Accuracy (Ngen=256)
Exact Match Accuracy (Ngen=512)
Average Exact Match Accuracy
DISPO
base_model=SFT w/ s1k,...
2026.02
75
82.3
83.2
80.2
DISPO
base_model=LLaDA-8B-In...
2026.02
74.2
80.6
81.8
78.9
diffu-GRPO
base_model=SFT w/ s1k,...
2026.02
73.2
81.1
82.1
78.8
diffu-GRPO
base_model=LLaDA-8B-In...
2026.02
72.6
79.8
81.9
78.1
LLaDA-8B-Instruct
base_model=LLaDA-8B-In...
2026.02
68.7
76.7
78.2
74.5
SFT w/ s1k
base_model=SFT w/ s1k
2026.02
66.5
78.8
81.1
75.5
Feedback
Search any
task
Search any
task