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

category chat · Qwen/Qwen3.8-27B · 105 runs in the model record
answer: 50.9 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
  Arc Pro B70 32GB                   | 4-bit | 44 tok/s       | measured | ttft 141ms | vram - GB | ctx 131,072
  CMP 170HX 64GB                     | 16-bit | 24.4 tok/s     | measured | ttft 5,898ms | vram - GB | ctx 65,799
  GB10 Grace Blackwell 256GB         | 16-bit | 8.2 tok/s      | reported | ttft 283ms | vram - GB | ctx 2,048
  H100 NVL 94GB ×2                   | 8-bit | 293.3 tok/s    | reported | ttft 69ms | vram - GB | ctx 2,048
  Radeon AI Pro R9700 32GB           | 4-bit | 285.5 tok/s    | measured | ttft 115ms | vram 19 GB | ctx 98,304
  Radeon AI Pro R9700 32GB           | 6-bit | 38.7 tok/s     | reported | ttft 77,056ms | vram 29.9 GB | ctx 131,072
  Radeon AI Pro R9700 32GB ×2        | 8-bit | 46.8 tok/s     | reported | ttft -ms | vram - GB | ctx 262,144
  Radeon AI Pro R9700 32GB ×3        | 4-bit | 28.4 tok/s     | measured | ttft 919ms | vram - GB | ctx 829
  Radeon AI Pro R9700 32GB ×3        | 8-bit | 52.9 tok/s     | measured | ttft 808ms | vram 48.6 GB | ctx 262,144
  RTX 3090 24GB                      | 4-bit | 113.2 tok/s    | reported | ttft 91ms | vram 22.4 GB | ctx 65,536
  RTX 3090 24GB                      | 6-bit | 53 tok/s       | reported | ttft 208ms | vram - GB | ctx 2,048
  RTX 3090 24GB ×2                   | 4-bit | 72.8 tok/s     | measured | ttft 949ms | vram - GB | ctx 262,144
  RTX 5060 Ti 16GB ×2                | 4-bit | 63.2 tok/s     | measured | ttft 190ms | vram - GB | ctx 131,072
  RTX 5090 32GB                      | 4-bit | 157.4 tok/s    | measured | ttft 279ms | vram 24 GB | ctx 170,000
  RTX PRO 6000 Blackwell 96GB        | 4-bit | 81.8 tok/s     | measured | ttft 215ms | vram - GB | ctx 2,048
  RTX PRO 6000 Blackwell 96GB        | 8-bit | 84.5 tok/s     | reported | ttft 379ms | vram - GB | ctx 2,048
  RTX PRO 6000 Blackwell 96GB        | 16-bit | 64.7 tok/s     | reported | ttft 92ms | vram - GB | ctx 262,144
  RX 7800 XT 16GB                    | 4-bit | 28 tok/s       | reported | ttft -ms | vram - GB | ctx 2,048
  RX 7900 XTX 24GB                   | 4-bit | 90.7 tok/s     | measured | ttft 646ms | vram 17.5 GB | ctx 32,768
  RX 7900 XTX 24GB ×2                | 8-bit | 35.2 tok/s     | measured | ttft 671ms | vram 31.7 GB | ctx 16,384
  Ryzen AI Max 395 128GB             | 4-bit | 65 tok/s       | measured | ttft 2,528ms | vram 18 GB | ctx 4,096
  Ryzen AI Max 395 128GB             | 8-bit | 14.2 tok/s     | measured | ttft 2,090ms | vram - GB | ctx 829
  Ryzen AI Max 395 64GB              | 4-bit | 18.5 tok/s     | measured | ttft 4,379ms | vram 17.4 GB | ctx 8,192
  Ryzen AI Max 395 64GB              | 6-bit | 17.3 tok/s     | measured | ttft 4,772ms | vram 25.7 GB | ctx 8,192
  Tesla V100 16GB ×4                 | 4-bit | 391 tok/s      | reported | ttft 349ms | vram 61 GB | ctx 32,768