Pair

GLM 5.3 Flash

Name: OpenWeights Terminal measured pairs
Creator: Dvidia
License: https://owterminal.com/cite

One quoted capture. Rank only lives inside a filter.

2x DGX Spark · NVFP4 · vLLM

Decode · tok/s

Prefill

—

tok/100W

6.3

decode, MTP off: 15 tok/s, flat to 128K

Stock vLLM NVFP4 on 2x Spark. Rank is MTP-off 15. Prefill range skipped.

Officialclaimed