OpenWeightsTerminal

Qwen3.8-27B

Qwen3.8-27B

RTX A5000 · NVFP4 · vLLM

Decode · tok/s
139.3
Prefill
—
tok/100W
60.6
na VLLM (v0.30.1rc1) ten sam model: 139.31 tokens/s

Same A5000 NVFP4 quote as the TensorFold row. Rank is the quoted 139.31.

Officialclaimed
Open pair