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 |
|---|---|---|---|---|
| Gemma 3 270M | Q4_K_M (4-bit) by unsloth | 52 | — | $1.38 |
| MiniCPM5 1B | Q4_K_M (4-bit) by openbmb | 22 | 8K | $3.29 |
| Llama 3.2 1B | IQ3_XS (3-bit) by maziyarpanahi | 18 | 4K | $4.01 |
| Gemma 4 E2B | Q3_K_S (3-bit) by unsloth | 13 | 4K | $5.60 |
| Phi 4 Mini-Reasoning | Q2_K (2-bit) by unsloth | 9.8 | — | $7.31 |
| Qwen 3 4B | IQ3_XXS (3-bit) by unsloth | 4.6 | 32K | $15.59 |
| Phi 4 Mini | IQ2_XS (2-bit) by mungert | 4.4 | — | $16.10 |
| Qwen3-VL-8B-Instruct-abliterated-v1 | Q2_K (2-bit) by mradermacher | 2.6 | 4K | $28.23 |