// RUN ON CONSUMER GPU

Run KoGPT 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 → ~451.5 yes 187k ~$1,999
NVIDIA GeForce RTX 4090 → ~254 yes 131k ~$1,599
NVIDIA GeForce RTX 3090 Ti → ~254 yes 131k ~$1,999
NVIDIA GeForce RTX 5080 → ~241.9 yes 75k ~$999
NVIDIA GeForce RTX 3090 → ~235.8 yes 131k ~$1,499
NVIDIA GeForce RTX 3080 Ti → ~229.8 yes 46k ~$1,199
NVIDIA GeForce RTX 5070 Ti → ~225.7 yes 75k ~$749
AMD Radeon RX 7900 XTX → ~214.1 yes 131k ~$999
NVIDIA GeForce RTX 3080 10GB → ~191.5 yes 32k ~$699
NVIDIA GeForce RTX 4080 → ~185.4 yes 75k ~$1,199
AMD Radeon RX 7900 XT → ~178.4 yes 103k ~$899
NVIDIA GeForce RTX 5070 → ~169.3 yes 46k ~$549
NVIDIA GeForce RTX 4070 Ti Super → ~169.3 yes 75k ~$799
NVIDIA GeForce RTX 3070 Ti → ~153.2 yes 18k ~$599
AMD Radeon RX 9070 XT → ~143.6 yes 75k ~$599
AMD Radeon RX 7800 XT → ~139.1 yes 75k ~$499
NVIDIA GeForce RTX 4070 Super → ~127 yes 46k ~$599
NVIDIA GeForce RTX 4070 → ~127 yes 46k ~$599
Intel Arc Pro B70 → ~126.8 yes 187k —
Intel Arc A770 16GB → ~116.8 yes 75k ~$329
AMD Radeon RX 6800 XT → ~114.2 yes 75k ~$649
AMD Radeon RX 6900 XT → ~114.2 yes 75k ~$999
NVIDIA GeForce RTX 3070 → ~112.9 yes 18k ~$499
NVIDIA GeForce RTX 5060 Ti 16GB → ~112.9 yes 75k ~$429
NVIDIA GeForce RTX 5060 Ti 8GB → ~112.9 yes 18k ~$379
Intel Arc Pro B60 → ~95.1 yes 131k —
Intel Arc B580 12GB → ~95.1 yes 46k ~$249
NVIDIA GeForce RTX 4060 Ti 8GB → ~72.6 yes 18k ~$399
NVIDIA GeForce RTX 4060 Ti 16GB → ~72.6 yes 75k ~$499
NVIDIA DGX Spark → ~68.8 yes 841k ~$3,999
AMD Ryzen AI Max+ 395 → ~57.1 yes 841k ~$1,999
Intel Arc Pro B50 → ~46.7 yes 75k —

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

→ KoGPT overview → All hardware

FAQ

What Consumer GPU hardware runs KoGPT?▶

On GenAIList, 32 Consumer GPU options can run KoGPT at Q4 — the fastest is the NVIDIA GeForce RTX 5090 at ~451.5 tok/s.

How fast is KoGPT on Consumer GPU hardware?▶

Estimated generation speed ranges from ~46.7 to ~451.5 tok/s at Q4 (memory-bandwidth bound). Click any option for its full estimate and any measured numbers.