Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Procedural reasoning on Procedural-skill v1.0 (test)
Loading...
95.5
Performance Score (BL)
pre-SFT Haiku-4-5
65.34
73.17
81
88.83
May 12, 2026
Performance Score (BL)
Consistency Score (CU)
Performance Delta
McNemar p-value
Delta 95% CI Lower Bound
Updated 2mo ago
Evaluation Results
Method
Method
Links
Performance Score (BL)
Consistency Score (CU)
Performance Delta
McNemar p-value
Delta 95% CI Lower Bound
pre-SFT Haiku-4-5
Framework=API, Model S...
2026.05
95.5
98.5
0.03
-
-
v2.0
Model Scale=4B, Traini...
2026.05
92
98.5
0.065
0.001
0.03
pre-SFT 4B
Framework=HF, Model Sc...
2026.05
85
84
0.01
-
-
v1.9
Model Scale=2B, Traini...
2026.05
81.5
91.5
0.1
0.0027
0.04
pre-SFT 2B
Framework=HF, Model Sc...
2026.05
75.5
81.5
0.06
-
-
v1
Model Scale=0.8B, Trai...
2026.05
71
70.5
0.005
-
-
pre-SFT 0.8B
Framework=HF, Model Sc...
2026.05
66.5
59
0.075
-
-
Feedback
Search any
task
Search any
task