{"model":{"id":"llama3.3-70b-q4","name":"Llama 3.3 70B Instruct","quantization":"Q4_K_M"},"hardware":{"slug":"rtx-3060","name":"NVIDIA GeForce RTX 3060","memoryGb":12},"verdict":"no","localVerdict":"local_unlikely","fitLevel":"Heavy","estimatedLoadGb":42,"usableMemoryGb":10.8,"estimatedTokensPerSec":0.7,"kvCacheByContext":[{"contextTokens":8192,"kvGb":2.5,"totalGb":44.5,"fits":false},{"contextTokens":16384,"kvGb":5,"totalGb":47,"fits":false},{"contextTokens":32768,"kvGb":10,"totalGb":52,"fits":false},{"contextTokens":65536,"kvGb":20,"totalGb":62,"fits":false},{"contextTokens":131072,"kvGb":40,"totalGb":82,"fits":false}],"upgradeGpu":{"slug":"rtx-6000-pro","name":"NVIDIA RTX PRO 6000 Blackwell"},"ollamaCommand":"ollama run llama3.3:70b-instruct-q4_K_M","reportUrl":"https://modelfit.io/can-i-run/llama3.3-70b-q4-on-rtx-3060/","source":"https://modelfit.io/","license":"CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/)","caveat":"Estimates from the ModelFit engine, not measurements."}