MTP results with vLLM inside

by unoid - opened Mar 3

Mar 3

Did this for 122b, now here 27B
vLLM Qwen3.5 27B Benchmark Results
FP8 Quantization - Speculative Decoding - March 3, 2026

RTX Pro 6000 Blackwell - 300W - vLLM 0.16.1rc1.dev173+g8fa68a8ce

unoid changed discussion status to closed Mar 3

Upload images, audio, and videos by dragging in the text input, pasting, or clicking here.

Tap or paste here to upload images

· Sign up or log in to comment