bestmodel.run / model — Qwen3.5-27B-GGUF (agent view)

category chat · unsloth/Qwen3.5-27B-GGUF · 12 runs in the model record
answer: 18.9 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 | 11.1 tok/s     | reported | ttft -ms | vram - GB | ctx 262,144
  Minisforum UM790 Pro 64GB          | 4-bit | 5 tok/s        | reported | ttft -ms | vram - GB | ctx 2,048
  Minisforum UM790 Pro 64GB          | 8-bit | 2.8 tok/s      | reported | ttft -ms | vram - GB | ctx 2,048
  Multi-GPU 48GB ×2                  | 8-bit | 27.7 tok/s     | reported | ttft -ms | vram - GB | ctx 2,048
  Radeon 890M 32GB                   | 4-bit | 4.2 tok/s      | reported | ttft 5,060ms | vram - GB | ctx 2,048
  Radeon AI Pro R9700 32GB           | 4-bit | 26.7 tok/s     | reported | ttft 466ms | vram - GB | ctx 2,048
  RTX 3090 24GB                      | 4-bit | 39.6 tok/s     | reported | ttft -ms | vram - GB | ctx 2,048
  RTX 4090 24GB                      | 4-bit | 45.4 tok/s     | reported | ttft -ms | vram - GB | ctx 2,048
  Ryzen AI Max 385 24GB              | 4-bit | 4.1 tok/s      | reported | ttft 5,060ms | vram - GB | ctx 2,048