Our new X account is live! Follow @wizwand_team for updates
Home
/
Benchmarks
Multilingual Mathematical Reasoning on MGSM 18 languages
Loading...
72.5
Accuracy
SP3F-7B
20.136
33.7305
47.325
60.9195
Jan 26, 2026
Accuracy
Language Fidelity
Updated 4d ago
Evaluation Results
Method
Method
Links
Accuracy
Language Fidelity
SP3F-7B
Training Stage=Full Pi...
2026.01
72.5
99.38
Qwen2.5-7B-Instruct
Training Stage=Instruct
2026.01
66.36
98.38
Qwen2.5-7B-Instruct + Translate Test
Training Stage=Instruc...
2026.01
66.15
95.81
Qwen2.5-7B + RLVR
Training Stage=SFT + RLVR
2026.01
65.34
99.75
Qwen2.5-7B + SFT
Training Stage=SFT
2026.01
33.66
91.37
Qwen2.5-7B
Training Stage=Base
2026.01
22.15
90.67
Feedback
Search any
task
Search any
task