List your company

Runtime efficiency on NIAH (500 samples)

34.01Latency (8K Context)H2O
33.639636.139838.6441.1402May 21, 2026

Evaluation results

MethodLinks
H2OInput Length=2K tokens, Output Length=1K tokens2026.0534.0134.7237.1540.3444.4653.37
SnapKVInput Length=2K tokens, Output Length=1K tokens2026.0534.2334.8637.3140.5644.6253.78
ZeroMergeInput Length=2K tokens, Output Length=1K tokens2026.0534.8735.5338.0741.2145.3554.62
Judge QInput Length=2K tokens, Output Length=1K tokens2026.0535.1235.8638.4341.7245.9455.29
Meta-SoftInput Length=2K tokens, Output Length=1K tokens2026.0535.2936.0338.7242.0546.3655.78
Full KVInput Length=2K tokens, Output Length=1K tokens2026.0543.2754.3879.93142.95271.83583.41