Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Mathematical Problem Solving on AIME 2025 (Accuracy avg@k)
Loading...
73.54
Accuracy (avg@k)
TRACE (FKL on Ky)
15.2584
30.3892
45.52
60.6508
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
73.54
TRACE (RKL on Ey)
Backbone=Qwen3-8B, Sym...
2026.05
70.42
GRPO
Backbone=Qwen3-8B, Sym...
2026.05
68.96
RLSD
Backbone=Qwen3-8B, Sym...
2026.05
68.54
Base
Backbone=Qwen3-8B, Sym...
2026.05
67.5
SRPO
Backbone=Qwen3-8B, Sym...
2026.05
54.17
SDPO
Backbone=Qwen3-8B, Sym...
2026.05
41.25
TRACE (RKL on Ey)
Backbone=Qwen3-1.7B, S...
2026.05
38.33
TRACE (FKL on Ky)
Backbone=Qwen3-1.7B, S...
2026.05
36.3
Base
Backbone=Qwen3-1.7B, S...
2026.05
35
GRPO
Backbone=Qwen3-1.7B, S...
2026.05
35
RLSD
Backbone=Qwen3-1.7B, S...
2026.05
32.92
SRPO
Backbone=Qwen3-1.7B, S...
2026.05
23.75
SDPO
Backbone=Qwen3-1.7B, S...
2026.05
17.5
Feedback
Search any
task
Search any
task