bestmodel.run / model — gemma-3-12b-it (agent view)

category chat · google/gemma-3-12b-it · 41 runs in the model record
answer: 44 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
  RTX 2080 8GB                       | 4-bit | 21 tok/s       | reported | ttft -ms | vram 8.1 GB | ctx 2,048
  RTX 3060 12GB                      | 4-bit | 40 tok/s       | reported | ttft -ms | vram 8.1 GB | ctx 2,048
  RTX 3060 12GB ×2                   | 4-bit | 34.3 tok/s     | measured | ttft 8,289ms | vram 12.2 GB | ctx 131,072
  RTX 3080 12GB                      | 4-bit | 51.5 tok/s     | measured | ttft -ms | vram 8.6 GB | ctx 2,048
  RTX 3090 24GB                      | 4-bit | 59.6 tok/s     | measured | ttft 3,320ms | vram - GB | ctx 131,072
  RTX3060+RTX2080-pooled-20GB 20GB   | 4-bit | 44 tok/s       | measured | ttft -ms | vram 8.1 GB | ctx 2,048
  Ryzen AI Max 395 64GB              | 4-bit | 24.6 tok/s     | reported | ttft 1,771ms | vram 9.1 GB | ctx 8,192