bestmodel.run / model — Qwen3.8-27B-NVFP4 (agent view)

category chat · unsloth/Qwen3.8-27B-NVFP4 · 15 runs in the model record
answer: 89.5 tok/s (model median)

Honesty ladder: measured > reported > extrapolated > formula > no data yet.
Empty fields stay empty — never filled with an estimate.

  rig | quant | median result | basis | ttft | peak vram | max context
  GB10 Grace Blackwell 128GB         | 4-bit | 19.1 tok/s     | measured | ttft 197ms | vram - GB | ctx 131,072
  GB10 Grace Blackwell 256GB         | 4-bit | 17.4 tok/s     | reported | ttft 168ms | vram - GB | ctx 2,048
  Radeon AI Pro R9700 32GB ×2        | 4-bit | 229.4 tok/s    | measured | ttft 68ms | vram 57 GB | ctx 262,144
  RTX 5090 32GB                      | 4-bit | 87.1 tok/s     | measured | ttft 135ms | vram 30 GB | ctx 153,664