Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Long-context retrieval and reasoning on RULER 16K
Loading...
85.2
RULER Score
Full Cache
57.432
64.641
71.85
79.059
Jun 1, 2026
RULER Score
Updated 1mo ago
Evaluation Results
Method
Method
Links
RULER Score
Full Cache
Backbone=LLaMA-3.1, Ca...
2026.06
85.2
Full Cache
Backbone=Qwen2-7B, Cac...
2026.06
80.5
MOMENTKV
Backbone=LLaMA-3.1, Ca...
2026.06
77.2
Ada-KV
Backbone=LLaMA-3.1, Ca...
2026.06
73.8
MOMENTKV
Backbone=Qwen2-7B, Cac...
2026.06
72
SnapKV
Backbone=LLaMA-3.1, Ca...
2026.06
71.5
PyramidKV
Backbone=LLaMA-3.1, Ca...
2026.06
70.2
Ada-KV
Backbone=Qwen2-7B, Cac...
2026.06
68.5
SnapKV
Backbone=Qwen2-7B, Cac...
2026.06
66.2
PyramidKV
Backbone=Qwen2-7B, Cac...
2026.06
65.5
H2O
Backbone=LLaMA-3.1, Ca...
2026.06
62.8
H2O
Backbone=Qwen2-7B, Cac...
2026.06
58.5
Feedback
Search any
task
Search any
task
