Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Long-context retrieval on RULER-NIAH 128k
Loading...
97.2
Accuracy
BF16
-3.888
22.356
48.6
74.844
May 18, 2026
Accuracy
Updated 14d ago
Evaluation Results
Method
Method
Links
Accuracy
BF16
Backbone=GLM-4.7-FP8†
2026.05
97.2
OSCAR
Backbone=GLM-4.7-FP8†
2026.05
97.2
QuaRot-INT2
Backbone=GLM-4.7-FP8†
2026.05
96.3
BF16
Backbone=Qwen3-4B-Thin...
2026.05
81
BF16
Backbone=Qwen3-8B
2026.05
78.2
OSCAR
Backbone=Qwen3-8B
2026.05
45
OSCAR
Backbone=Qwen3-4B-Thin...
2026.05
39.5
QuaRot-INT2
Backbone=Qwen3-4B-Thin...
2026.05
0
QuaRot-INT2
Backbone=Qwen3-8B
2026.05
0
Feedback
Search any
task
Search any
task