bestmodel.run / model — Qwen3.8-27B-FP8 (agent view)

category chat · Qwen/Qwen3.8-27B-FP8 · 24 runs in the model record
answer: 80.8 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
  Arc Pro B70 32GB                   | 8-bit | 54.2 tok/s     | measured | ttft 177ms | vram - GB | ctx 32,768
  Arc Pro B70 32GB ×2                | 8-bit | 64.3 tok/s     | measured | ttft 94ms | vram - GB | ctx 33,024
  B200 192GB                         | 8-bit | 106.2 tok/s    | reported | ttft 37ms | vram 158.6 GB | ctx 32,768
  GB10 Grace Blackwell 128GB         | 8-bit | 14.8 tok/s     | reported | ttft -ms | vram - GB | ctx 65,536
  H200 SXM 141GB                     | 8-bit | 252.4 tok/s    | measured | ttft 45ms | vram 111.4 GB | ctx 8,192
  RTX 5090 32GB                      | 8-bit | 16.8 tok/s     | reported | ttft 144ms | vram 30.9 GB | ctx 8,192
  RTX PRO 6000 Blackwell 96GB        | 8-bit | 100.4 tok/s    | reported | ttft 68ms | vram - GB | ctx 262,144