// RUN ON CONSUMER GPU

Run Gemma 3n on Consumer GPU hardware

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

HardwareEst tok/sFits 8k?Max ctx~Price
NVIDIA GeForce RTX 5090 → ~354.6 yes 180k ~$1,999
NVIDIA GeForce RTX 4090 → ~199.5 yes 123k ~$1,599
NVIDIA GeForce RTX 3090 Ti → ~199.5 yes 123k ~$1,999
NVIDIA GeForce RTX 5080 → ~190 yes 67k ~$999
NVIDIA GeForce RTX 3090 → ~185.2 yes 123k ~$1,499
NVIDIA GeForce RTX 3080 Ti → ~180.5 yes 39k ~$1,199
NVIDIA GeForce RTX 5070 Ti → ~177.3 yes 67k ~$749
AMD Radeon RX 7900 XTX → ~168.2 yes 123k ~$999
NVIDIA GeForce RTX 3080 10GB → ~150.4 yes 25k ~$699
NVIDIA GeForce RTX 4080 → ~145.7 yes 67k ~$1,199
AMD Radeon RX 7900 XT → ~140.1 yes 95k ~$899
NVIDIA GeForce RTX 5070 → ~133 yes 39k ~$549
NVIDIA GeForce RTX 4070 Ti Super → ~133 yes 67k ~$799
NVIDIA GeForce RTX 3070 Ti → ~120.3 yes 11k ~$599
AMD Radeon RX 9070 XT → ~112.8 yes 67k ~$599
AMD Radeon RX 7800 XT → ~109.3 yes 67k ~$499
NVIDIA GeForce RTX 4070 Super → ~99.7 yes 39k ~$599
NVIDIA GeForce RTX 4070 → ~99.7 yes 39k ~$599
Intel Arc Pro B70 → ~99.6 yes 180k —
Intel Arc A770 16GB → ~91.7 yes 67k ~$329
AMD Radeon RX 6800 XT → ~89.7 yes 67k ~$649
AMD Radeon RX 6900 XT → ~89.7 yes 67k ~$999
NVIDIA GeForce RTX 3070 → ~88.7 yes 11k ~$499
NVIDIA GeForce RTX 5060 Ti 16GB → ~88.7 yes 67k ~$429
NVIDIA GeForce RTX 5060 Ti 8GB → ~88.7 yes 11k ~$379
Intel Arc Pro B60 → ~74.7 yes 123k —
Intel Arc B580 12GB → ~74.7 yes 39k ~$249
NVIDIA GeForce RTX 4060 Ti 8GB → ~57 yes 11k ~$399
NVIDIA GeForce RTX 4060 Ti 16GB → ~57 yes 67k ~$499
NVIDIA DGX Spark → ~54 yes 834k ~$3,999
AMD Ryzen AI Max+ 395 → ~44.8 yes 834k ~$1,999
Intel Arc Pro B50 → ~36.7 yes 67k —

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

→ Gemma 3n overview → All hardware

FAQ

What Consumer GPU hardware runs Gemma 3n?▶

On GenAIList, 32 Consumer GPU options can run Gemma 3n at Q4 — the fastest is the NVIDIA GeForce RTX 5090 at ~354.6 tok/s.

How fast is Gemma 3n on Consumer GPU hardware?▶

Estimated generation speed ranges from ~36.7 to ~354.6 tok/s at Q4 (memory-bandwidth bound). Click any option for its full estimate and any measured numbers.