Gemma on AMD RDNA3

Can I run Gemma 3 12B on an Radeon RX 7900 XTX?

Yes. Gemma 3 12B at Q4_K_M uses 8.6 GB of the 21.8 GB available on a Radeon RX 7900 XTX, and runs at about 71.1 tokens per second. You can push the context to 128K.

Runs comfortably

8.6 GB of 21.8 GB · 39%
022 GB
Weights 7.0 GB
KV cache 0.8 GB
Runtime overhead 0.8 GB

Gemma 3 12B at Q4_K_M leaves 13.2 GB spare on a Radeon RX 7900 XTX. There is room to raise the context length or move up a quantisation level.

Generation71.1tok/s
Prompt processing2117tok/s
Max context128Ktokens
KV per 1K tokens0GB

Every quantisation of Gemma 3 12B on a Radeon RX 7900 XTX

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 — 8.49
INT8 / W8A8 8.50 11.8 GB 13.5 GB Runs comfortably 128K 43.5
Q8_0 (GGUF) 8.50 11.8 GB 13.5 GB Runs comfortably 128K 43.5
FP8 (E4M3) 8.00 11.4 GB 13.0 GB Runs comfortably 128K 45.2
Q6_K 6.56 9.3 GB 10.9 GB Runs comfortably 128K 54.4
Q5_K_M 5.67 8.1 GB 9.7 GB Runs comfortably 128K 62.3
Q5_K_S 5.52 7.8 GB 9.5 GB Runs comfortably 128K 63.8
AWQ 4-bit 4.25 7.4 GB 9.0 GB Runs comfortably 128K 67.2
GPTQ 4-bit 4.25 7.4 GB 9.0 GB Runs comfortably 128K 67.2
MXFP4 4.25 7.4 GB 9.0 GB Runs comfortably 128K 67.2
Q4_K_M 4.85 7.0 GB 8.6 GB Runs comfortably 128K 71.1
Q4_K_S 4.58 6.6 GB 8.2 GB Runs comfortably 128K 74.5
Q4_0 4.55 6.6 GB 8.2 GB Runs comfortably 128K 74.9
IQ4_XS 4.25 6.2 GB 7.8 GB Runs comfortably 128K 79.1
Q3_K_M 3.91 5.7 GB 7.4 GB Runs comfortably 128K 84.6
IQ3_M 3.70 5.5 GB 7.1 GB Runs comfortably 128K 88.3
IQ3_XXS 3.06 4.6 GB 6.3 GB Runs comfortably 128K 102
Q2_K 2.63 4.1 GB 5.7 GB Runs comfortably 128K 114
IQ2_XXS 2.06 3.3 GB 5.0 GB Runs comfortably 128K 135
IQ1_M 1.75 2.9 GB 4.6 GB Runs comfortably 128K 150

Also worth checking

Gemma 3 12B on every GPU
Full hardware tableEverything that runs on a Radeon RX 7900 XTX
Full model tableWhy the KV cache is the part that bites
Guide