Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Mathematical Reasoning on MATH 500 (M500 Score)
Loading...
95.8
M500 Score
GEAR
87.792
89.871
91.95
94.029
May 12, 2026
M500 Score
Updated 21d ago
Evaluation Results
Method
Method
Links
M500 Score
GEAR
Backbone=Qwen3-4B, Eva...
2026.05
95.8
GEAR
Backbone=Qwen3-4B, Eva...
2026.05
95.2
OPSD+RL
Backbone=Qwen3-4B, Eva...
2026.05
94.9
GEAR
Backbone=Qwen3-8B, Eva...
2026.05
94.7
OPSD
Backbone=Qwen3-4B, Eva...
2026.05
93.7
GRPO
Backbone=Qwen3-4B, Eva...
2026.05
93.3
ARPO
Backbone=Qwen3-4B, Eva...
2026.05
93.2
GRPO
Backbone=Qwen3-8B, Eva...
2026.05
93.2
MT-GRPO
Backbone=Qwen3-4B, Eva...
2026.05
93.1
OPSD
Backbone=Qwen3-8B, Eva...
2026.05
93.1
ARPO
Backbone=Qwen3-8B, Eva...
2026.05
92.8
MT-GRPO
Backbone=Qwen3-8B, Eva...
2026.05
92.7
OPSD+RL
Backbone=Qwen3-8B, Eva...
2026.05
92.5
GRPO
Backbone=Qwen3-4B, Eva...
2026.05
92.4
Base
Backbone=Qwen3-8B, Eva...
2026.05
89.6
Base
Backbone=Qwen3-4B, Eva...
2026.05
88.1
Feedback
Search any
task
Search any
task