Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Mathematical Problem Solving on AIME 2024 (Accuracy (avg@k))
Loading...
80.21
Accuracy (avg@k)
TRACE (FKL on Ky)
13.6916
30.9608
48.23
65.4992
May 11, 2026
Accuracy (avg@k)
Updated 2mo ago
Evaluation Results
Method
Method
Links
Accuracy (avg@k)
TRACE (FKL on Ky)
Backbone=Qwen3-8B, Sym...
2026.05
80.21
TRACE (RKL on Ey)
Backbone=Qwen3-8B, Sym...
2026.05
78.54
GRPO
Backbone=Qwen3-8B, Sym...
2026.05
77.08
Base
Backbone=Qwen3-8B, Sym...
2026.05
76.25
RLSD
Backbone=Qwen3-8B, Sym...
2026.05
75.42
SRPO
Backbone=Qwen3-8B, Sym...
2026.05
69.17
SDPO
Backbone=Qwen3-8B, Sym...
2026.05
52.5
RLSD
Backbone=Qwen3-1.7B, S...
2026.05
45.42
TRACE (RKL on Ey)
Backbone=Qwen3-1.7B, S...
2026.05
44.92
TRACE (FKL on Ky)
Backbone=Qwen3-1.7B, S...
2026.05
44.83
GRPO
Backbone=Qwen3-1.7B, S...
2026.05
44.58
Base
Backbone=Qwen3-1.7B, S...
2026.05
43.75
SRPO
Backbone=Qwen3-1.7B, S...
2026.05
24.17
SDPO
Backbone=Qwen3-1.7B, S...
2026.05
16.25
Feedback
Search any
task
Search any
task