Qwen on NVIDIA Ampere

Can I run Qwen3 4B on an RTX 3060 12GB?

Yes. Qwen3 4B at Q4_K_M uses 4.2 GB of the 10.5 GB available on a RTX 3060 12GB, and runs at about 77.4 tokens per second. You can push the context to 53K.

Runs comfortably

4.2 GB of 10.5 GB · 40%
010 GB
Weights 2.3 GB
KV cache 1.1 GB
Runtime overhead 0.8 GB

Qwen3 4B at Q4_K_M leaves 6.3 GB spare on a RTX 3060 12GB. There is room to raise the context length or move up a quantisation level.

Generation77.4tok/s
Prompt processing1306tok/s
Max context53Ktokens
KV per 1K tokens0GB

Every quantisation of Qwen3 4B 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 7.5 GB 9.4 GB Runs comfortably 16K 27.8
INT8 / W8A8 8.50 3.9 GB 5.8 GB Runs comfortably 41K 50.1
Q8_0 (GGUF) 8.50 3.9 GB 5.8 GB Runs comfortably 41K 50.1
FP8 (E4M3) 8.00 3.7 GB 5.7 GB Runs comfortably 42K 51.7
Q6_K 6.56 3.1 GB 5.0 GB Runs comfortably 47K 61.2
Q5_K_M 5.67 2.7 GB 4.6 GB Runs comfortably 50K 69.0
Q5_K_S 5.52 2.6 GB 4.5 GB Runs comfortably 51K 70.5
AWQ 4-bit 4.25 2.5 GB 4.4 GB Runs comfortably 51K 71.9
GPTQ 4-bit 4.25 2.5 GB 4.4 GB Runs comfortably 51K 71.9
MXFP4 4.25 2.5 GB 4.4 GB Runs comfortably 51K 71.9
Q4_K_M 4.85 2.3 GB 4.2 GB Runs comfortably 53K 77.4
Q4_K_S 4.58 2.2 GB 4.1 GB Runs comfortably 53K 80.6
Q4_0 4.55 2.2 GB 4.1 GB Runs comfortably 53K 81.0
IQ4_XS 4.25 2.0 GB 4.0 GB Runs comfortably 54K 84.8
Q3_K_M 3.91 1.9 GB 3.8 GB Runs comfortably 55K 89.7
IQ3_M 3.70 1.8 GB 3.7 GB Runs comfortably 56K 93.0
IQ3_XXS 3.06 1.5 GB 3.5 GB Runs comfortably 58K 105
Q2_K 2.63 1.4 GB 3.3 GB Runs comfortably 59K 114
IQ2_XXS 2.06 1.1 GB 3.0 GB Runs comfortably 61K 130
IQ1_M 1.75 1.0 GB 2.9 GB Runs comfortably 62K 141

Also worth checking

Qwen3 4B 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