// RUN ON CONSUMER GPU

Run CPM-Bee on Consumer GPU hardware

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

HardwareEst tok/sFits 8k?Max ctx~Price
NVIDIA GeForce RTX 5090 → ~278.4 yes 136k ~$1,999
NVIDIA GeForce RTX 4090 → ~156.6 yes 91k ~$1,599
NVIDIA GeForce RTX 3090 Ti → ~156.6 yes 91k ~$1,999
NVIDIA GeForce RTX 5080 → ~149.1 yes 47k ~$999
NVIDIA GeForce RTX 3090 → ~145.4 yes 91k ~$1,499
NVIDIA GeForce RTX 3080 Ti → ~141.7 yes 24k ~$1,199
NVIDIA GeForce RTX 5070 Ti → ~139.2 yes 47k ~$749
AMD Radeon RX 7900 XTX → ~132 yes 91k ~$999
NVIDIA GeForce RTX 3080 10GB → ~118.1 yes 13k ~$699
NVIDIA GeForce RTX 4080 → ~114.3 yes 47k ~$1,199
AMD Radeon RX 7900 XT → ~110 yes 69k ~$899
NVIDIA GeForce RTX 5070 → ~104.4 yes 24k ~$549
NVIDIA GeForce RTX 4070 Ti Super → ~104.4 yes 47k ~$799
AMD Radeon RX 9070 XT → ~88.6 yes 47k ~$599
AMD Radeon RX 7800 XT → ~85.8 yes 47k ~$499
NVIDIA GeForce RTX 4070 Super → ~78.3 yes 24k ~$599
NVIDIA GeForce RTX 4070 → ~78.3 yes 24k ~$599
Intel Arc Pro B70 → ~78.2 yes 136k —
Intel Arc A770 16GB → ~72 yes 47k ~$329
AMD Radeon RX 6800 XT → ~70.4 yes 47k ~$649
AMD Radeon RX 6900 XT → ~70.4 yes 47k ~$999
NVIDIA GeForce RTX 5060 Ti 16GB → ~69.6 yes 47k ~$429
Intel Arc Pro B60 → ~58.6 yes 91k —
Intel Arc B580 12GB → ~58.6 yes 24k ~$249
NVIDIA GeForce RTX 4060 Ti 16GB → ~44.7 yes 47k ~$499
NVIDIA DGX Spark → ~42.4 yes 660k ~$3,999
AMD Ryzen AI Max+ 395 → ~35.2 yes 660k ~$1,999
Intel Arc Pro B50 → ~28.8 yes 47k —

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

→ CPM-Bee overview → All hardware

FAQ

What Consumer GPU hardware runs CPM-Bee?▶

On GenAIList, 28 Consumer GPU options can run CPM-Bee at Q4 — the fastest is the NVIDIA GeForce RTX 5090 at ~278.4 tok/s.

How fast is CPM-Bee on Consumer GPU hardware?▶

Estimated generation speed ranges from ~28.8 to ~278.4 tok/s at Q4 (memory-bandwidth bound). Click any option for its full estimate and any measured numbers.