Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Attention Latency on Qwen3-30B-A3B-Instruct-2507 (32K Context)
Loading...
1
Speedup Factor
Dense
0.7608
0.8229
0.885
0.9471
May 16, 2026
Speedup Factor
Updated 2mo ago
Evaluation Results
Method
Method
Links
Speedup Factor
Dense
GPU=H200 SXM, TP=2, Ba...
2026.05
1
CompactAttention-FP
GPU=H200 SXM, TP=2, Ba...
2026.05
0.83
QUOKA
GPU=H200 SXM, TP=2, Ba...
2026.05
0.81
FlashPrefill
GPU=H200 SXM, TP=2, Ba...
2026.05
0.77
Feedback
Search any
task
Search any
task