NVIDIA RTX 4070

12 GB VRAM200 W TDP2 providers3 offerings2 regions

The NVIDIA RTX 4070 is a consumer-grade GPU built on the Ada Lovelace architecture, featuring 12 GB of memory with 504 GB/s of memory bandwidth and a 200 W TDP. It delivers 29.1 TFLOPS of FP32 performance. Currently available from 2 providers starting at $0.12/GPU/hour, with a market median of $0.18/GPU/hour across 3 configurations.

Cheapest

$0.120/hr

Theta EdgeCloud

Median

$0.180/hr

across 3 offerings

Most expensive

$0.200/hr

Theta EdgeCloud

90-day trend

-21.7%

median rate

Hardware specifications

Same across all providers

Memory

12 GB

GDDR6X

What LLMs fit? ↓

TDP

200 W

Memory bandwidth

504 GB/s

Architecture

Ada Lovelace

FP32 performance

29.1 TFLOPS

Brand

NVIDIA

Series

GeForce RTX

Market price history

Median price across all providers

Prices updated 43 minutes ago

2 offerings from 2 providers

Sorted by price ascending. Compare configurations side-by-side.

Provider	Count	vCPU	RAM	Per GPU hour	Total/hr	Action
Vast.aiPromoted Affiliate promotion — we may earn a commission.	×1	5	21 GB	$0.180	$0.18	Launch
Theta EdgeCloudTrendingCheapest	×1	22	8 GB	From$0.120 +14.3% 30d	From$0.12	Launch
Vast.aiReferral linkTrending	×1	5	21 GB	$0.180 +63.6% 30d	$0.18	Launch

LLMs that fit on the NVIDIA RTX 4070

Estimated VRAM for popular open-weight LLMs at 16-bit (FP16/BF16), 8-bit (FP8/INT8) and 4-bit (INT4/MXFP4) precision, against 12 GB per GPU.

Model	Parameters	16-bit	8-bit	4-bit
Gemma 3 12B Google	12.2B	29 GB Too large	16 GB Too large	9.4 GB 1× · 2.6 GB free
Granite 4.1 8B IBM	8.8B	21 GB Too large	12 GB Barely fits on 1×	6.8 GB 1× · 5.2 GB free
Qwen3 8B Alibaba	8.2B	20 GB Too large	11 GB 1× · 1.2 GB free	6.3 GB 1× · 5.7 GB free
Llama 3.1 8B Instruct Meta	8B	19 GB Too large	11 GB 1× · 1.4 GB free	6.2 GB 1× · 5.8 GB free
Mistral 7B Instruct v0.3 Mistral AI	7.2B	17 GB Too large	9.6 GB 1× · 2.4 GB free	5.6 GB 1× · 6.4 GB free
Qwen3 4B Alibaba	4B	9.7 GB 1× · 2.3 GB free	5.3 GB 1× · 6.7 GB free	3.1 GB 1× · 8.9 GB free

Estimates: model weights plus 20% for KV cache and runtime overhead, at short context lengths. Long contexts and large batches need more. Rows marked “tight” hold the weights but not that full margin. A ×N figure is the total VRAM across N of these GPUs and makes no claim about interconnect throughput. How we estimate this · Start from a model instead

Want more from this page?

See incorrect data?

Frequently asked questions

How much does the RTX 4070 cost per hour?

The RTX 4070 starts at $0.12 per GPU-hour, with a median of $0.18 across 3 available configurations. Actual pricing depends on the provider, region, and whether you rent on-demand, spot or reserved.

Which cloud providers offer the RTX 4070?

2 providers currently list the RTX 4070: Theta EdgeCloud, Vast.ai. Compare their hourly pricing, regions and machine specs in the table above.

How much VRAM does the RTX 4070 have?

The RTX 4070 comes with 12 GB of VRAM, which determines the largest models and batch sizes it can hold in memory for training and inference.

Which LLMs can I run on the RTX 4070?

With 12 GB of VRAM, a single RTX 4070 can serve models such as Gemma 3 12B, Granite 4.1 8B, Qwen3 8B, Llama 3.1 8B Instruct. The table above lists the estimated VRAM for each model at 16-bit, 8-bit and 4-bit precision, including the multi-GPU configurations needed for larger models.

Where is the RTX 4070 cheapest?

The lowest on-demand price we currently track for the RTX 4070 is $0.12 per GPU-hour at Theta EdgeCloud. Because providers update pricing and availability frequently, check the live table above before you launch.