RTX 3060 12 GB
4 records across 1 models (quants: IQ4_XS).
| hardware | model | quant | backend | tok/s | pp tok/s | W | ttft s | ctx/tg | date | class | flags | source |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| RTX 3060 12 GB | Mistral Small 4 119B | IQ4_XS | llama.cpp (CUDA) | 19.36 | – | – | – | – | 2026-10-03 | community | – | source |
| RTX 3060 12 GB | Mistral Small 4 119B | IQ4_XS | llama.cpp (CUDA) | 18.68 | – | – | – | – | 2026-10-03 | community | – | source |
| RTX 3060 12 GB | Mistral Small 4 119B | IQ4_XS | llama.cpp (CUDA) | 5.40 | – | – | – | – | 2026-10-03 | community | – | source |
| RTX 3060 12 GB | Mistral Small 4 119B | IQ4_XS | llama.cpp (CUDA) | 4.83 | – | – | – | – | 2026-10-03 | community | – | source |
All rows cite their source. Provenance: community 40018 sourced 1085