Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Candidate-ranking Retrieval on Diagnostic Probes 1024-token context
Loading...
96.7
Needle Accuracy
Full
40.748
55.274
69.8
84.326
Jun 1, 2026
Needle Accuracy
Bridge-window Accuracy
Semantic Exact Accuracy
Semantic Paraphrase Accuracy
Clean Semantic Average
Semantic Distractor Accuracy
Semantic Boundary Accuracy
Updated 1mo ago
Evaluation Results
Method
Method
Links
Needle Accuracy
Bridge-window Accuracy
Semantic Exact Accuracy
Semantic Paraphrase Accuracy
Clean Semantic Average
Semantic Distractor Accuracy
Semantic Boundary Accuracy
Full
Architecture=Full
2026.06
96.7
98.9
98.5
90.6
94.6
31.4
54.9
SWA
Architecture=SWA
2026.06
51.3
99.1
54.1
44.4
49.3
28.2
27.8
PBB
Architecture=PBB
2026.06
43.8
84.2
48.1
42
45.1
23.7
29.2
Bridge
Architecture=Bridge
2026.06
43.7
81.2
46.2
37.8
42
23.5
27.6
SE-Bridge
Architecture=SE-Bridge
2026.06
43.6
98.3
56
47.1
51.6
22.3
29
Block
Architecture=Block
2026.06
42.9
41.5
44.9
33.4
39.2
20.4
27.1
Feedback
Search any
task
Search any
task
