// RUN ON CONSUMER GPU

Run Kosmos-2 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 → ~1,740 yes 413k ~$1,999
NVIDIA GeForce RTX 4090 → ~978.8 yes 300k ~$1,599
NVIDIA GeForce RTX 3090 Ti → ~978.8 yes 300k ~$1,999
NVIDIA GeForce RTX 5080 → ~932.1 yes 188k ~$999
NVIDIA GeForce RTX 3090 → ~908.8 yes 300k ~$1,499
NVIDIA GeForce RTX 3080 Ti → ~885.5 yes 132k ~$1,199
NVIDIA GeForce RTX 5070 Ti → ~870 yes 188k ~$749
AMD Radeon RX 7900 XTX → ~825 yes 300k ~$999
NVIDIA GeForce RTX 3080 10GB → ~737.9 yes 104k ~$699
NVIDIA GeForce RTX 4080 → ~714.6 yes 188k ~$1,199
AMD Radeon RX 7900 XT → ~687.5 yes 244k ~$899
NVIDIA GeForce RTX 5070 → ~652.5 yes 132k ~$549
NVIDIA GeForce RTX 4070 Ti Super → ~652.5 yes 188k ~$799
NVIDIA GeForce RTX 3070 Ti → ~590.4 yes 76k ~$599
AMD Radeon RX 9070 XT → ~553.4 yes 188k ~$599
AMD Radeon RX 7800 XT → ~536.3 yes 188k ~$499
NVIDIA GeForce RTX 4070 Super → ~489.4 yes 132k ~$599
NVIDIA GeForce RTX 4070 → ~489.4 yes 132k ~$599
Intel Arc Pro B70 → ~488.6 yes 413k —
Intel Arc A770 16GB → ~450 yes 188k ~$329
AMD Radeon RX 6800 XT → ~440 yes 188k ~$649
AMD Radeon RX 6900 XT → ~440 yes 188k ~$999
NVIDIA GeForce RTX 3070 → ~435 yes 76k ~$499
NVIDIA GeForce RTX 5060 Ti 16GB → ~435 yes 188k ~$429
NVIDIA GeForce RTX 5060 Ti 8GB → ~435 yes 76k ~$379
Intel Arc Pro B60 → ~366.4 yes 300k —
Intel Arc B580 12GB → ~366.4 yes 132k ~$249
NVIDIA GeForce RTX 4060 Ti 8GB → ~279.6 yes 76k ~$399
NVIDIA GeForce RTX 4060 Ti 16GB → ~279.6 yes 188k ~$499
NVIDIA DGX Spark → ~265.1 yes 1721k ~$3,999
AMD Ryzen AI Max+ 395 → ~220 yes 1721k ~$1,999
Intel Arc Pro B50 → ~180 yes 188k —
NVIDIA GeForce RTX 2050 Mobile → ~108.8 yes 20k —

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

→ Kosmos-2 overview → All hardware

FAQ

What Consumer GPU hardware runs Kosmos-2?▶

On GenAIList, 33 Consumer GPU options can run Kosmos-2 at Q4 — the fastest is the NVIDIA GeForce RTX 5090 at ~1,740 tok/s.

How fast is Kosmos-2 on Consumer GPU hardware?▶

Estimated generation speed ranges from ~108.8 to ~1,740 tok/s at Q4 (memory-bandwidth bound). Click any option for its full estimate and any measured numbers.