Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Skill Retrieval on SkillResolve-Bench 1.0 (held-out pairs)
Loading...
91.1
R@20
SkillResolve
77.892
81.321
84.75
88.179
Jun 9, 2026
R@20
NDCG@20
HSR@20
Updated 1mo ago
Evaluation Results
Method
Method
Links
R@20
NDCG@20
HSR@20
SkillResolve
K=20
2026.06
91.1
74.7
0
BGE rerank top20
K=20
2026.06
87.7
48.2
87.9
Hybrid lexical RRF
K=20
2026.06
87.7
58.7
87.9
Attribution-listwise
K=20
2026.06
85.3
65.7
0
SkillRouter
K=20
2026.06
78.4
57.7
81.4
Feedback
Search any
task
Search any
task