Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Mathematical Problem-solving on MATH (exact match %)
Loading...
60.12
Exact Match (%)
DS2-INSTRUCT
-0.928
14.921
30.77
46.619
Mar 13, 2026
Exact Match (%)
Updated 1mo ago
Evaluation Results
Method
Method
Links
Exact Match (%)
DS2-INSTRUCT
Model Family=Qwen2.5
2026.03
60.12
InstructMix
Model Family=Qwen2.5
2026.03
56.24
Self-Instruct
Model Family=Qwen2.5
2026.03
53.67
ExploreInstruct
Model Family=Qwen2.5
2026.03
51.38
Zero-Shot
Model Family=Qwen2.5
2026.03
49.44
DS2-INSTRUCT
Model Family=Llama3
2026.03
23.16
DS2-INSTRUCT
Model Family=Mistral
2026.03
21.82
InstructMix
Model Family=Llama3
2026.03
14.92
ExploreInstruct
Model Family=Llama3
2026.03
13.06
ExploreInstruct
Model Family=Mistral
2026.03
11.74
Self-Instruct
Model Family=Llama3
2026.03
11.47
InstructMix
Model Family=Mistral
2026.03
8.67
Self-Instruct
Model Family=Mistral
2026.03
5.94
Zero-Shot
Model Family=Llama3
2026.03
3.45
Zero-Shot
Model Family=Mistral
2026.03
1.42
Feedback
Search any
task
Search any
task