// RUN ON CONSUMER GPU

Run ProGen2-base 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 → ~3,644 yes 420k ~$1,999
NVIDIA GeForce RTX 4090 → ~2,049.7 yes 307k ~$1,599
NVIDIA GeForce RTX 3090 Ti → ~2,049.7 yes 307k ~$1,999
NVIDIA GeForce RTX 5080 → ~1,952.1 yes 195k ~$999
NVIDIA GeForce RTX 3090 → ~1,903.3 yes 307k ~$1,499
NVIDIA GeForce RTX 3080 Ti → ~1,854.5 yes 139k ~$1,199
NVIDIA GeForce RTX 5070 Ti → ~1,822 yes 195k ~$749
AMD Radeon RX 7900 XTX → ~1,727.7 yes 307k ~$999
NVIDIA GeForce RTX 3080 10GB → ~1,545.4 yes 111k ~$699
NVIDIA GeForce RTX 4080 → ~1,496.6 yes 195k ~$1,199
AMD Radeon RX 7900 XT → ~1,439.8 yes 251k ~$899
NVIDIA GeForce RTX 5070 → ~1,366.5 yes 139k ~$549
NVIDIA GeForce RTX 4070 Ti Super → ~1,366.5 yes 195k ~$799
NVIDIA GeForce RTX 3070 Ti → ~1,236.4 yes 83k ~$599
AMD Radeon RX 9070 XT → ~1,159 yes 195k ~$599
AMD Radeon RX 7800 XT → ~1,123 yes 195k ~$499
NVIDIA GeForce RTX 4070 Super → ~1,024.9 yes 139k ~$599
NVIDIA GeForce RTX 4070 → ~1,024.9 yes 139k ~$599
Intel Arc Pro B70 → ~1,023.2 yes 420k —
Intel Arc A770 16GB → ~942.4 yes 195k ~$329
AMD Radeon RX 6800 XT → ~921.5 yes 195k ~$649
AMD Radeon RX 6900 XT → ~921.5 yes 195k ~$999
NVIDIA GeForce RTX 3070 → ~911 yes 83k ~$499
NVIDIA GeForce RTX 5060 Ti 16GB → ~911 yes 195k ~$429
NVIDIA GeForce RTX 5060 Ti 8GB → ~911 yes 83k ~$379
Intel Arc Pro B60 → ~767.4 yes 307k —
Intel Arc B580 12GB → ~767.4 yes 139k ~$249
NVIDIA GeForce RTX 4060 Ti 8GB → ~585.6 yes 83k ~$399
NVIDIA GeForce RTX 4060 Ti 16GB → ~585.6 yes 195k ~$499
NVIDIA DGX Spark → ~555.1 yes 1728k ~$3,999
AMD Ryzen AI Max+ 395 → ~460.7 yes 1728k ~$1,999
Intel Arc Pro B50 → ~377 yes 195k —
NVIDIA GeForce RTX 2050 Mobile → ~227.7 yes 27k —

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

→ ProGen2-base overview → All hardware

FAQ

What Consumer GPU hardware runs ProGen2-base?▶

On GenAIList, 33 Consumer GPU options can run ProGen2-base at Q4 — the fastest is the NVIDIA GeForce RTX 5090 at ~3,644 tok/s.

How fast is ProGen2-base on Consumer GPU hardware?▶

Estimated generation speed ranges from ~227.7 to ~3,644 tok/s at Q4 (memory-bandwidth bound). Click any option for its full estimate and any measured numbers.