Qwen on NVIDIA Ampere

Can I run Qwen3 8B on an RTX 3080 10GB?

Yes. Qwen3 8B at Q4_K_M uses 6.7 GB of the 8.6 GB available on a RTX 3080 10GB, and runs at about 88.4 tokens per second. You can push the context to 22K.

Runs comfortably

6.7 GB of 8.6 GB · 77%
09 GB
Weights 4.7 GB
KV cache 1.1 GB
Runtime overhead 0.8 GB

Qwen3 8B at Q4_K_M leaves 1.9 GB spare on a RTX 3080 10GB. There is room to raise the context length or move up a quantisation level.

Generation88.4tok/s
Prompt processing1538tok/s
Max context22Ktokens
KV per 1K tokens0GB

Every quantisation of Qwen3 8B on a RTX 3080 10GB

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

QuantisationbpwWeights TotalVerdictMax ctxtok/s
FP16 / BF16 16.00 15.3 GB 17.2 GB Won't fit — 4.08
INT8 / W8A8 8.50 7.8 GB 9.8 GB Won't fit — 20.0
Q8_0 (GGUF) 8.50 7.8 GB 9.8 GB Won't fit — 20.0
FP8 (E4M3) 8.00 7.6 GB 9.6 GB Won't fit 1K 23.0
Q6_K 6.56 6.3 GB 8.2 GB Fits, but tight 11K 68.7
AWQ 4-bit 4.25 5.8 GB 7.7 GB Runs comfortably 14K 74.1
GPTQ 4-bit 4.25 5.8 GB 7.7 GB Runs comfortably 14K 74.1
MXFP4 4.25 5.8 GB 7.7 GB Runs comfortably 14K 74.1
Q5_K_M 5.67 5.4 GB 7.4 GB Runs comfortably 17K 78.4
Q5_K_S 5.52 5.3 GB 7.2 GB Runs comfortably 18K 80.2
Q4_K_M 4.85 4.7 GB 6.7 GB Runs comfortably 22K 88.4
Q4_K_S 4.58 4.5 GB 6.4 GB Runs comfortably 23K 92.1
Q4_0 4.55 4.5 GB 6.4 GB Runs comfortably 23K 92.5
IQ4_XS 4.25 4.2 GB 6.2 GB Runs comfortably 25K 97.1
Q3_K_M 3.91 4.0 GB 5.9 GB Runs comfortably 27K 103
IQ3_M 3.70 3.8 GB 5.7 GB Runs comfortably 28K 107
IQ3_XXS 3.06 3.3 GB 5.2 GB Runs comfortably 32K 121
Q2_K 2.63 2.9 GB 4.9 GB Runs comfortably 35K 133
IQ2_XXS 2.06 2.5 GB 4.4 GB Runs comfortably 38K 152
IQ1_M 1.75 2.2 GB 4.2 GB Runs comfortably 40K 165

Also worth checking

Qwen3 8B on every GPU
Full hardware tableEverything that runs on a RTX 3080 10GB
Full model tableWhy the KV cache is the part that bites
Guide