bestmodel.run / model — Qwen3.6-27B-FP8 (agent view)

category chat · Qwen/Qwen3.6-27B-FP8 · 15 runs in the model record
answer: 72.5 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
  Arc Pro B70 32GB ×2                | 8-bit | 20.1 tok/s     | reported | ttft -ms | vram - GB | ctx 1,024
  GB10 Grace Blackwell 128GB         | 8-bit | 9.5 tok/s      | reported | ttft -ms | vram - GB | ctx 65,536
  RTX 3090 24GB ×2                   | 8-bit | 46.3 tok/s     | measured | ttft 421ms | vram 22 GB | ctx 262,144
  RTX 3090 Ti 24GB ×2                | 8-bit | 70.3 tok/s     | reported | ttft 465ms | vram 22 GB | ctx 131,072
  RTX 5090 32GB ×2                   | 8-bit | 179.9 tok/s    | reported | ttft 2,222ms | vram 58.3 GB | ctx 262,144
  RTX PRO 6000 Blackwell 96GB        | 8-bit | 92.6 tok/s     | measured | ttft 222ms | vram 96 GB | ctx 131,072
  RTX PRO 6000 Blackwell 96GB        | 16-bit | 96.2 tok/s     | reported | ttft 72ms | vram - GB | ctx 262,144
  RTX PRO 6000 Blackwell Workstation Edition 96GB | 8-bit | 24.8 tok/s     | reported | ttft 9,998ms | vram 82.4 GB | ctx 65,536