bestmodel.run / model — gemma-4-31B-it-GGUF (agent view)

category chat · unsloth/gemma-4-31B-it-GGUF · 26 runs in the model record
answer: 15.5 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
  M5 Max 128GB                       | 4-bit | 24.1 tok/s     | reported | ttft 565ms | vram - GB | ctx 2,048
  M5 Max 128GB                       | 6-bit | 18.1 tok/s     | reported | ttft 678ms | vram - GB | ctx 2,048
  M5 Max 128GB                       | 8-bit | 15.3 tok/s     | reported | ttft 660ms | vram - GB | ctx 2,048
  M5 Max 128GB                       | 16-bit | 8.9 tok/s      | reported | ttft 858ms | vram - GB | ctx 2,048
  Minisforum UM790 Pro 64GB          | 4-bit | 4 tok/s        | reported | ttft -ms | vram - GB | ctx 2,048
  Radeon AI Pro R9700 32GB           | 4-bit | 25.4 tok/s     | reported | ttft 555ms | vram - GB | ctx 2,048
  RTX 3060 12GB ×2                   | 4-bit | 15.2 tok/s     | measured | ttft 10,911ms | vram 21.5 GB | ctx 98,304
  RTX 3090 24GB                      | 4-bit | 38.8 tok/s     | reported | ttft -ms | vram - GB | ctx 2,048
  RTX 3090 24GB ×2                   | 6-bit | 14.4 tok/s     | reported | ttft 2,263ms | vram 20.7 GB | ctx 32,768
  RTX 5090 32GB                      | 6-bit | 51.7 tok/s     | reported | ttft 155ms | vram - GB | ctx 2,048