Inference · hardware

What runs on Windows, 8 CPUs · 8 GB.

Every open model a machine of this kind has run on the network, at its fastest build, fastest first. Speed is tokens a second, measured and signed by the machine itself.

2models measured
23fastest tok/s
1machine
ModelFastest build heretok/sContext$ / M tokens
Sarvam-1Q4_K_M (4-bit) by bartowski238K$7.51
Gemma 4 E2BQ4_K_M (4-bit) by unsloth9.9128K$17.92