Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Reasoning on AIME (pass@1 accuracy)
Loading...
73.33
Pass@1 Accuracy
Baseline
38.6668
47.6659
56.665
65.6641
Oct 1, 2025
Pass@1 Accuracy
Updated 23d ago
Evaluation Results
Method
Method
Links
Pass@1 Accuracy
Baseline
Model=QwQ-32B, Bit-Wid...
2025.10
73.33
ThinKV (k=1024)
Model=QwQ-32B, Bit-Wid...
2025.10
70.28
PM-KVQ
Model=QwQ-32B, Bit-Wid...
2025.10
67.86
KIVI
Model=QwQ-32B, Bit-Wid...
2025.10
60.56
Baseline
Model=R1-Qwen-14B, Bit...
2025.10
53.33
ThinKV (k=1024)
Model=R1-Qwen-14B, Bit...
2025.10
50
PM-KVQ
Model=R1-Qwen-14B, Bit...
2025.10
43.33
KIVI
Model=R1-Qwen-14B, Bit...
2025.10
40
Feedback
Search any
task
Search any
task