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.
| Model | Fastest build here | tok/s | Context | $ / M tokens |
|---|---|---|---|---|
| Llama 3.2 3B | IQ1_S (1-bit) by unsloth | 225 | — | $1.14 |
| Phi 4 Mini-Reasoning | IQ3_XXS (3-bit) by unsloth | 189 | 4K | $1.36 |
| Gemma 4 E2B | Q4_K_M (4-bit) by google | 170 | 128K | $1.54 |
| Phi 4 Mini | Q2_K (2-bit) by maziyarpanahi | 154 | — | $1.66 |
| Qwen 3 4B | IQ1_S (1-bit) by unsloth | 90 | — | $2.84 |