// RUN ON CONSUMER GPU

Run Llama 3.2 3B on Consumer GPU hardware

33 options ranked by estimated speed at Q4 / 8k context.

HardwareEst tok/sFits 8k?Max ctx~Price
NVIDIA GeForce RTX 5090 → ~867.3 yes 399k ~$1,999
NVIDIA GeForce RTX 4090 → ~487.9 yes 287k ~$1,599
NVIDIA GeForce RTX 3090 Ti → ~487.9 yes 287k ~$1,999
NVIDIA GeForce RTX 5080 → ~464.6 yes 174k ~$999
NVIDIA GeForce RTX 3090 → ~453 yes 287k ~$1,499
NVIDIA GeForce RTX 3080 Ti → ~441.4 yes 118k ~$1,199
NVIDIA GeForce RTX 5070 Ti → ~433.6 yes 174k ~$749
AMD Radeon RX 7900 XTX → ~411.2 yes 287k ~$999
NVIDIA GeForce RTX 3080 10GB → ~367.8 yes 90k ~$699
NVIDIA GeForce RTX 4080 → ~356.2 yes 174k ~$1,199
AMD Radeon RX 7900 XT → ~342.7 yes 230k ~$899
NVIDIA GeForce RTX 5070 → ~325.2 yes 118k ~$549
NVIDIA GeForce RTX 4070 Ti Super → ~325.2 yes 174k ~$799
NVIDIA GeForce RTX 3070 Ti → ~294.3 yes 62k ~$599
AMD Radeon RX 9070 XT → ~275.9 yes 174k ~$599
AMD Radeon RX 7800 XT → ~267.3 yes 174k ~$499
NVIDIA GeForce RTX 4070 Super → ~243.9 yes 118k ~$599
NVIDIA GeForce RTX 4070 → ~243.9 yes 118k ~$599
Intel Arc Pro B70 → ~243.5 yes 399k —
Intel Arc A770 16GB → ~224.3 yes 174k ~$329
AMD Radeon RX 6800 XT → ~219.3 yes 174k ~$649
AMD Radeon RX 6900 XT → ~219.3 yes 174k ~$999
NVIDIA GeForce RTX 3070 → ~216.8 yes 62k ~$499
NVIDIA GeForce RTX 5060 Ti 16GB → ~216.8 yes 174k ~$429
NVIDIA GeForce RTX 5060 Ti 8GB → ~216.8 yes 62k ~$379
Intel Arc Pro B60 → ~182.6 yes 287k —
Intel Arc B580 12GB → ~182.6 yes 118k ~$249
NVIDIA GeForce RTX 4060 Ti 8GB → ~139.4 yes 62k ~$399
NVIDIA GeForce RTX 4060 Ti 16GB → ~139.4 yes 174k ~$499
NVIDIA DGX Spark → ~132.1 yes 1707k ~$3,999
AMD Ryzen AI Max+ 395 → ~109.7 yes 1707k ~$1,999
Intel Arc Pro B50 → ~89.7 yes 174k —
NVIDIA GeForce RTX 2050 Mobile → ~54.2 tight 6k —

Estimate at Q4 (memory-bound). Click any option for its full per-hardware estimate and measured numbers.

→ Llama 3.2 3B overview → All hardware

FAQ

What Consumer GPU hardware runs Llama 3.2 3B?▶

On GenAIList, 33 Consumer GPU options can run Llama 3.2 3B at Q4 — the fastest is the NVIDIA GeForce RTX 5090 at ~867.3 tok/s.

How fast is Llama 3.2 3B on Consumer GPU hardware?▶

Estimated generation speed ranges from ~54.2 to ~867.3 tok/s at Q4 (memory-bandwidth bound). Click any option for its full estimate and any measured numbers.