bestmodel.run / model — gemma-4-26B-A4B-it-GGUF (agent view)

category chat · unsloth/gemma-4-26B-A4B-it-GGUF · 69 runs in the model record
answer: 90.3 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                   | 8-bit | 97.6 tok/s     | measured | ttft 602ms | vram - GB | ctx 32,768
  GB10 Grace Blackwell 128GB         | 4-bit | 50.7 tok/s     | reported | ttft -ms | vram - GB | ctx 262,144
  GB10 Grace Blackwell 128GB         | 8-bit | 38.8 tok/s     | reported | ttft -ms | vram - GB | ctx 262,144
  M5 Max 128GB                       | 4-bit | 99.2 tok/s     | reported | ttft 124ms | vram - GB | ctx 2,048
  M5 Pro 24GB                        | 4-bit | 55 tok/s       | reported | ttft 446ms | vram - GB | ctx 4,224
  Minisforum UM790 Pro 64GB          | 4-bit | 19.5 tok/s     | reported | ttft -ms | vram - GB | ctx 2,048
  Radeon RX 7900 XTX 24GB            | 4-bit | 94.3 tok/s     | reported | ttft -ms | vram 15.3 GB | ctx 640
  RTX 3060 12GB ×2                   | 4-bit | 68.2 tok/s     | measured | ttft 2,495ms | vram 18.6 GB | ctx 131,072
  RTX 5060 Ti 16GB                   | 3-bit | 103.4 tok/s    | reported | ttft 148ms | vram - GB | ctx 2,048
  RTX 5090 32GB                      | 3-bit | 255.7 tok/s    | reported | ttft 41ms | vram - GB | ctx 2,048
  RTX 5090 32GB                      | 4-bit | 395 tok/s      | reported | ttft 274ms | vram 18.7 GB | ctx 32,768
  Ryzen AI Max 395 64GB              | 4-bit | 49.2 tok/s     | reported | ttft 1,004ms | vram 17.8 GB | ctx 8,192