Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Long-Context Retrieval on RULER 64K
Loading...
97.33
Score
HiLS-Attn HoPE-Q-Cal
-3.8932
22.3859
48.665
74.9441
Apr 27, 2026
May 8, 2026
May 19, 2026
May 30, 2026
Jun 10, 2026
Jun 21, 2026
Jul 3, 2026
Score
Updated 18d ago
Evaluation Results
Method
Method
Links
Score
HiLS-Attn HoPE-Q-Cal
Tuning Protocol=Full p...
2026.07
97.33
Nemotron-3-Ultra 550B-A55B-Base
shots=0-shot
2026.06
95.3
Kimi-K2 Base
shots=0-shot
2026.06
93.79
DeepSeek-V3.2 Exp-Base
shots=0-shot
2026.06
93.3
Mistral-Large-3 675B-Base-2512
shots=0-shot
2026.06
90.11
Hybrid-KSA
setting=Train-from-scr...
2026.04
69.95
Hybrid-GDN
setting=Train-from-scr...
2026.04
69.39
Full
setting=Train-from-scr...
2026.04
65.32
Hybrid-SCA
setting=Train-from-scr...
2026.04
63.54
Hybrid-SWA
setting=Train-from-scr...
2026.04
63.03
KSA
setting=Train-from-scr...
2026.04
57.13
GLM-4.5 Base
shots=0-shot
2026.06
16.12
HiLS-Attn RoPE-Q-Cal
Tuning Protocol=Full p...
2026.07
4.33
Olmo3 512swa
Tuning Protocol=Full p...
2026.07
1.33
LMK token tuning
Tuning Protocol=Freezi...
2026.07
1
HiLS-Attn NoPE-Q-Cal
Tuning Protocol=Full p...
2026.07
0.33
Olmo3 Base
Tuning Protocol=Freezi...
2026.07
0
Feedback
Search any
task
Search any
task