// RUN ON CONSUMER GPU

Run Spec-Drafter 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 → ~5,568 yes 422k ~$1,999
NVIDIA GeForce RTX 4090 → ~3,132 yes 310k ~$1,599
NVIDIA GeForce RTX 3090 Ti → ~3,132 yes 310k ~$1,999
NVIDIA GeForce RTX 5080 → ~2,982.9 yes 197k ~$999
NVIDIA GeForce RTX 3090 → ~2,908.3 yes 310k ~$1,499
NVIDIA GeForce RTX 3080 Ti → ~2,833.7 yes 141k ~$1,199
NVIDIA GeForce RTX 5070 Ti → ~2,784 yes 197k ~$749
AMD Radeon RX 7900 XTX → ~2,640 yes 310k ~$999
NVIDIA GeForce RTX 3080 10GB → ~2,361.4 yes 113k ~$699
NVIDIA GeForce RTX 4080 → ~2,286.9 yes 197k ~$1,199
AMD Radeon RX 7900 XT → ~2,200 yes 254k ~$899
NVIDIA GeForce RTX 5070 → ~2,088 yes 141k ~$549
NVIDIA GeForce RTX 4070 Ti Super → ~2,088 yes 197k ~$799
NVIDIA GeForce RTX 3070 Ti → ~1,889.1 yes 85k ~$599
AMD Radeon RX 9070 XT → ~1,771 yes 197k ~$599
AMD Radeon RX 7800 XT → ~1,716 yes 197k ~$499
NVIDIA GeForce RTX 4070 Super → ~1,566 yes 141k ~$599
NVIDIA GeForce RTX 4070 → ~1,566 yes 141k ~$599
Intel Arc Pro B70 → ~1,563.4 yes 422k —
Intel Arc A770 16GB → ~1,440 yes 197k ~$329
AMD Radeon RX 6800 XT → ~1,408 yes 197k ~$649
AMD Radeon RX 6900 XT → ~1,408 yes 197k ~$999
NVIDIA GeForce RTX 3070 → ~1,392 yes 85k ~$499
NVIDIA GeForce RTX 5060 Ti 16GB → ~1,392 yes 197k ~$429
NVIDIA GeForce RTX 5060 Ti 8GB → ~1,392 yes 85k ~$379
Intel Arc Pro B60 → ~1,172.6 yes 310k —
Intel Arc B580 12GB → ~1,172.6 yes 141k ~$249
NVIDIA GeForce RTX 4060 Ti 8GB → ~894.9 yes 85k ~$399
NVIDIA GeForce RTX 4060 Ti 16GB → ~894.9 yes 197k ~$499
NVIDIA DGX Spark → ~848.3 yes 1731k ~$3,999
AMD Ryzen AI Max+ 395 → ~704 yes 1731k ~$1,999
Intel Arc Pro B50 → ~576 yes 197k —
NVIDIA GeForce RTX 2050 Mobile → ~348 yes 29k —

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

→ Spec-Drafter overview → All hardware

FAQ

What Consumer GPU hardware runs Spec-Drafter?▶

On GenAIList, 33 Consumer GPU options can run Spec-Drafter at Q4 — the fastest is the NVIDIA GeForce RTX 5090 at ~5,568 tok/s.

How fast is Spec-Drafter on Consumer GPU hardware?▶

Estimated generation speed ranges from ~348 to ~5,568 tok/s at Q4 (memory-bandwidth bound). Click any option for its full estimate and any measured numbers.