{"model":{"id":"gemma4-12b-q8","name":"Gemma 4 12B (Q8)","quantization":"Q8_0"},"hardware":{"slug":"rtx-3090","name":"NVIDIA GeForce RTX 3090","memoryGb":24},"verdict":"yes","localVerdict":"local_feasible","fitLevel":"Excellent","estimatedLoadGb":12.8,"usableMemoryGb":21.6,"estimatedTokensPerSec":38.2,"kvCacheByContext":[{"contextTokens":8192,"kvGb":1.5,"totalGb":14.3,"fits":true},{"contextTokens":16384,"kvGb":3,"totalGb":15.8,"fits":true},{"contextTokens":32768,"kvGb":6,"totalGb":18.8,"fits":true},{"contextTokens":65536,"kvGb":12,"totalGb":24.8,"fits":false},{"contextTokens":131072,"kvGb":24,"totalGb":36.8,"fits":false}],"upgradeGpu":null,"ollamaCommand":"ollama run gemma4:12b-it-q8_0","reportUrl":"https://modelfit.io/can-i-run/gemma4-12b-q8-on-rtx-3090/","source":"https://modelfit.io/","license":"CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/)","caveat":"Estimates from the ModelFit engine, not measurements."}