Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Simple Retrieval on Ruler S-NIAH
Loading...
100
Accuracy
GPT-5
68.8
76.9
85
93.1
Mar 3, 2026
Accuracy
Updated 1mo ago
Evaluation Results
Method
Method
Links
Accuracy
GPT-5
Architecture=Base LLM
2026.03
100
RLM
Backbone=GPT-5, Recurs...
2026.03
100
DeepSeek v3.2
Architecture=Base LLM
2026.03
100
Kimi K2
Architecture=Base LLM
2026.03
100
RLM
Backbone=Kimi K2, Recu...
2026.03
90
RLM
Backbone=Kimi K2, Recu...
2026.03
90
RLM
Backbone=DeepSeek v3.2...
2026.03
85
RLM
Backbone=DeepSeek v3.2...
2026.03
70
Feedback
Search any
task
Search any
task
