Skip to content

NVIDIA · Ada Lovelace · 2023

NVIDIA RTX 4070 Ti. Not in stock right now.

12 GB of GDDR6X and 80.2 TFLOPS FP16 non-tensor, for training and inference. The RTX 4090 is on demand now from $0.58/hr per GPU on an 8-GPU machine.

Not in stock right now

RTX 4070 Ti is not in stock right now.

Leave your work email to register interest. These requests decide which GPUs we bring on demand next.

The card

12 GB of GDDR6X. At 504 GB/s.

Memory
12 GB
GDDR6X
Bandwidth
504 GB/s
Memory to compute
FP16
80.2
non-tensor TFLOPS
Per node
2 GPUs
PCIe, 285 W each

Spec sheet

How the RTX 4070 Ti compares. Against its closest alternatives.

FP16 non-tensor

NVIDIA RTX 4070 Ti80.2 TFLOPS
NVIDIA RTX 309071 TFLOPS
NVIDIA RTX 408097.5 TFLOPS

Memory

NVIDIA RTX 4070 Ti12 GB
NVIDIA RTX 309024 GB
NVIDIA RTX 408016 GB
NVIDIA V10032 GB

Bandwidth

NVIDIA RTX 4070 Ti504 GB/s
NVIDIA RTX 3090936 GB/s
NVIDIA RTX 4080736 GB/s
NVIDIA V100900 GB/s

Alternatives

Alternatives to the RTX 4070 Ti. Same account, different price and performance.

GPUMemory · FP16 non-tensorStands outPrice gapFrom
NVIDIA RTX 4070 TiThis GPU12 GB GDDR6X · 80.2 TFLOPSNot in stock right now
NVIDIA RTX 309024 GB GDDR6X · 71 TFLOPS+12 GB memory$0.41/hr
NVIDIA RTX 408016 GB GDDR6X · 97.5 TFLOPS+4 GB memoryNot in stock right now
NVIDIA V10032 GB HBM2+20 GB memory$3.35/hr
NVIDIA A4048 GB GDDR6+36 GB memory$2.66/hr

Full specs

NVIDIA RTX 4070 Ti specifications.

Memory

Memory
12 GB
Type
GDDR6X
Bandwidth
504 GB/s

Compute

FP32
40.1 TFLOPS
FP16 non-tensor
80.2 TFLOPS
INT8
321 TOPS
CUDA cores
7,680

Power and form

TDP
285 W
Form factor
Consumer
Architecture
Ada Lovelace
Released
2023

Node

GPUs per node
2
Interconnect
PCIe
Maker
NVIDIA

Questions

Renting the RTX 4070 Ti. What buyers ask first.

Other GPUs

Browse every GPU by architecture.