Gemma on NVIDIA Ampere

Can I run Gemma 3 12B on an RTX 3060 12GB?

Yes. Gemma 3 12B at Q4_K_M uses 8.6 GB of the 10.5 GB available on a RTX 3060 12GB, and runs at about 29.7 tokens per second. You can push the context to 38K.

Runs comfortably

8.6 GB of 10.5 GB · 82%
010 GB
Weights 7.0 GB
KV cache 0.8 GB
Runtime overhead 0.8 GB

Gemma 3 12B at Q4_K_M leaves 1.9 GB spare on a RTX 3060 12GB. There is room to raise the context length or move up a quantisation level.

Generation29.7tok/s
Prompt processing430tok/s
Max context38Ktokens
KV per 1K tokens0GB

Every quantisation of Gemma 3 12B on a RTX 3060 12GB

Highlighted row is the highest quality that still fits at 8K context.

QuantisationbpwWeights TotalVerdictMax ctxtok/s
FP16 / BF16 16.00 22.7 GB 24.4 GB Won't fit — 2.48
INT8 / W8A8 8.50 11.8 GB 13.5 GB Won't fit — 8.03
Q8_0 (GGUF) 8.50 11.8 GB 13.5 GB Won't fit — 8.03
FP8 (E4M3) 8.00 11.4 GB 13.0 GB Won't fit — 9.13
Q6_K 6.56 9.3 GB 10.9 GB Won't fit 896 17.6
Q5_K_M 5.67 8.1 GB 9.7 GB Fits, but tight 21K 26.0
Q5_K_S 5.52 7.8 GB 9.5 GB Fits, but tight 24K 26.6
AWQ 4-bit 4.25 7.4 GB 9.0 GB Runs comfortably 31K 28.0
GPTQ 4-bit 4.25 7.4 GB 9.0 GB Runs comfortably 31K 28.0
MXFP4 4.25 7.4 GB 9.0 GB Runs comfortably 31K 28.0
Q4_K_M 4.85 7.0 GB 8.6 GB Runs comfortably 38K 29.7
Q4_K_S 4.58 6.6 GB 8.2 GB Runs comfortably 44K 31.1
Q4_0 4.55 6.6 GB 8.2 GB Runs comfortably 44K 31.3
IQ4_XS 4.25 6.2 GB 7.8 GB Runs comfortably 51K 33.1
Q3_K_M 3.91 5.7 GB 7.4 GB Runs comfortably 58K 35.4
IQ3_M 3.70 5.5 GB 7.1 GB Runs comfortably 62K 37.0
IQ3_XXS 3.06 4.6 GB 6.3 GB Runs comfortably 75K 42.8
Q2_K 2.63 4.1 GB 5.7 GB Runs comfortably 84K 48.0
IQ2_XXS 2.06 3.3 GB 5.0 GB Runs comfortably 96K 57.0
IQ1_M 1.75 2.9 GB 4.6 GB Runs comfortably 103K 63.4

Also worth checking

Gemma 3 12B on every GPU
Full hardware tableEverything that runs on a RTX 3060 12GB
Full model tableWhy the KV cache is the part that bites
Guide