// RUN ON CONSUMER GPU

Run MSA Transformer 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 → ~27,840 yes 425k ~$1,999
NVIDIA GeForce RTX 4090 → ~15,660 yes 313k ~$1,599
NVIDIA GeForce RTX 3090 Ti → ~15,660 yes 313k ~$1,999
NVIDIA GeForce RTX 5080 → ~14,914.3 yes 201k ~$999
NVIDIA GeForce RTX 3090 → ~14,541.4 yes 313k ~$1,499
NVIDIA GeForce RTX 3080 Ti → ~14,168.6 yes 145k ~$1,199
NVIDIA GeForce RTX 5070 Ti → ~13,920 yes 201k ~$749
AMD Radeon RX 7900 XTX → ~13,200 yes 313k ~$999
NVIDIA GeForce RTX 3080 10GB → ~11,807.1 yes 117k ~$699
NVIDIA GeForce RTX 4080 → ~11,434.3 yes 201k ~$1,199
AMD Radeon RX 7900 XT → ~11,000 yes 257k ~$899
NVIDIA GeForce RTX 5070 → ~10,440 yes 145k ~$549
NVIDIA GeForce RTX 4070 Ti Super → ~10,440 yes 201k ~$799
NVIDIA GeForce RTX 3070 Ti → ~9,445.7 yes 89k ~$599
AMD Radeon RX 9070 XT → ~8,855 yes 201k ~$599
AMD Radeon RX 7800 XT → ~8,580 yes 201k ~$499
NVIDIA GeForce RTX 4070 Super → ~7,830 yes 145k ~$599
NVIDIA GeForce RTX 4070 → ~7,830 yes 145k ~$599
Intel Arc Pro B70 → ~7,817.1 yes 425k —
Intel Arc A770 16GB → ~7,200 yes 201k ~$329
AMD Radeon RX 6800 XT → ~7,040 yes 201k ~$649
AMD Radeon RX 6900 XT → ~7,040 yes 201k ~$999
NVIDIA GeForce RTX 3070 → ~6,960 yes 89k ~$499
NVIDIA GeForce RTX 5060 Ti 16GB → ~6,960 yes 201k ~$429
NVIDIA GeForce RTX 5060 Ti 8GB → ~6,960 yes 89k ~$379
Intel Arc Pro B60 → ~5,862.9 yes 313k —
Intel Arc B580 12GB → ~5,862.9 yes 145k ~$249
NVIDIA GeForce RTX 4060 Ti 8GB → ~4,474.3 yes 89k ~$399
NVIDIA GeForce RTX 4060 Ti 16GB → ~4,474.3 yes 201k ~$499
NVIDIA DGX Spark → ~4,241.3 yes 1734k ~$3,999
AMD Ryzen AI Max+ 395 → ~3,520 yes 1734k ~$1,999
Intel Arc Pro B50 → ~2,880 yes 201k —
NVIDIA GeForce RTX 2050 Mobile → ~1,740 yes 32k —

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

→ MSA Transformer overview → All hardware

FAQ

What Consumer GPU hardware runs MSA Transformer?▶

On GenAIList, 33 Consumer GPU options can run MSA Transformer at Q4 — the fastest is the NVIDIA GeForce RTX 5090 at ~27,840 tok/s.

How fast is MSA Transformer on Consumer GPU hardware?▶

Estimated generation speed ranges from ~1,740 to ~27,840 tok/s at Q4 (memory-bandwidth bound). Click any option for its full estimate and any measured numbers.