Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Inference Efficiency on 128K-context
Loading...
101
TTFT
IT-SPEED-16
10.52
34.01
57.5
80.99
May 7, 2026
TTFT
TPOT
KV Cache Utilization
Updated 26d ago
Evaluation Results
Method
Method
Links
TTFT
TPOT
KV Cache Utilization
IT-SPEED-16
K=16, BoS=false
2026.05
101
55
50
IT-SPEED-16+BoS
K=16, BoS=true
2026.05
101
55
50
IT-SPEED-20
K=20, BoS=false
2026.05
60
36
37.5
IT-SPEED-20+BoS
K=20, BoS=true
2026.05
60
36
37.5
IT-SPEED-24
K=24, BoS=false
2026.05
33
22
25
IT-SPEED-24+BoS
K=24, BoS=true
2026.05
33
22
25
IT-SPEED-28
K=28, BoS=false
2026.05
14
10
12.5
IT-SPEED-28+BoS
K=28, BoS=true
2026.05
14
10
12.5
Feedback
Search any
task
Search any
task