How much does a RTX 4090 24 GB rental cost?

From about 25 ₽ per GPU-hour at the lowest current offer. Prices change with availability, so the live figure above is the one to trust.

How is billing calculated?

Per minute while the server runs, from your balance. The price is fixed at launch. A server that fails to start is not billed.

Do I pay for the GPU when the server is stopped?

No GPU time is billed while the server is stopped. The disk is still billed, so keep only the data you need on it.

Data on the disk stays while the server exists. Deleting the server deletes its disk too, so download results first.

Which template should I start with?

Ollama for chatting with language models, ComfyUI or Forge for images and video, PyTorch with Jupyter for training and experiments, Ubuntu with CUDA when you want to install everything yourself.

GPU availability updates every minute

gpu.nz

GPU Catalog Pricing Templates Regions Docs

en ru

Launch GPU

Home GPUsRTX 4090 24 GB

NVIDIAAda Lovelace

Rent RTX 4090 24 GB

Name: RTX 4090 24 GB
Brand: NVIDIA
Availability: InStock

Лучшая цена за производительность для Stable Diffusion, FLUX и LoRA.

from25₽per GPU per hour

Prices now

Minimum: 25₽
Median: 80₽
Maximum: 4,531₽
Servers free: 396
GPUs free: 768
Countries: 40

Architecture: Ada Lovelace
VRAM per GPU: 24 GB
Class: Игровая

Prices now

gpu.nz

Cloud GPUs by the minute.

t.me/gpunz

Product

GPU catalog
Templates
Prices by model
Regions
Pricing
Solutions
Comparisons
Guides

Company

Business
Docs
Support
Telegram channel

Legal

Public offer
Privacy policy
Personal data consent
Acceptable use
3D model credits

Rent RTX 4090 24 GB — from 25 ₽/h · gpu.nz

en ru

Home GPUsRTX 4090 24 GB

NVIDIAAda Lovelace

Rent RTX 4090 24 GB

Лучшая цена за производительность для Stable Diffusion, FLUX и LoRA.

from25₽per GPU per hour

Prices now

Minimum: 25₽
Median: 80₽
Maximum: 4,531₽
Servers free: 396
GPUs free: 768
Countries: 40

Architecture: Ada Lovelace
VRAM per GPU: 24 GB
Class: Игровая

Specifications

Video memory: 24 GB
Memory type: GDDR6X
Architecture: Ada Lovelace
Price from: 25₽/h

What it is good for

LLM (int8 / int4)

8Bfp16 · Fits
14Bint8 · Fits
32Bint4 · Fits
70BNot enough memory

What it is good for

Stable Diffusion 1.5Fits
SDXL: generationFits
FLUX.1 dev in fp8Fits
FLUX.1 dev in bf16Fits
Video: Wan 2.1 1.3BFits
Video: Wan 14B, HunyuanVideofrom 48 GB
SDXL LoRA trainingFits
QLoRA for 7B (4-bit)Fits
LoRA for 7B (fp16)Fits
Full fine-tuning of 7Bfrom 112 GB

Estimates for short context and typical settings. See the guide on memory for LLMs. How much VRAM an LLM needs: a table and a rule of thumb

Rental price

On-demand	per hour	per day
1 GPU	25₽	598₽
2 GPU	111₽	2,662₽
4 GPU	229₽	5,494₽
8 GPU	528₽	12,670₽

Interruptible: 15₽/h per GPUInterruptible bid price per GPU. The server can be stopped if someone outbids you.

Prices and availability come from the live catalog and update automatically.

Where it is available now

How to start

Pick a server with the memory you need and a price you like in the catalog. Memory and type filters are on the catalog page.
Launch it with a template: Ollama for chatting with LLMs, ComfyUI or Forge for images and video, PyTorch + Jupyter for experiments.
Work in the browser or over SSH. Stop the server when you are done: GPU minutes stop being billed.

Pricing

Solutions for this GPU

Run LLMs on a rented GPU: Llama, Qwen, DeepSeek, Mistral
Stable Diffusion and FLUX on a rented GPU: Forge and SDXL online
ComfyUI online on your own server: FLUX, Wan and HunyuanVideo
Fine-tuning and LoRA training on a rented GPU
JupyterLab with a GPU for rent: PyTorch, CUDA and data science
GPU rendering on a rented server: Blender Cycles and 3D by the minute
Video generation on a rented GPU: Wan and HunyuanVideo in ComfyUI

Prices now

Filters

Available onlyBusy servers are shown greyed and cannot be rented

GPU count

VRAM per GPU

Price, ₽/h per GPU

full range

4–9059

Region

USUnited States457
NONorway179
OMOman139
CACanada119
TWTaiwan113
JPJapan87

Reliability

Verified hosts only

More specs

CPU cores, minRAM, GB, minDisk, GB, minDownload, Mbit/s, minCUDA, minTFLOPS, min

Found: 1,877 serversAvailable: 396

Sort

GPU	Server	Region	Reliab.	Rent
RTX 4090 24 GB876 GB/s	24 vCPU · 118 GB RAM394 GB · ↓ 471	MYMalaysia	93.3%	Busy
RTX 4090 24 GB879 GB/s	20 vCPU · 125 GB RAM608 GB · ↓ 453	INIndia	99.7%	Busy
RTX 4090 24 GB899 GB/s	24 vCPU · 47 GB RAM295 GB · ↓ 1,183	PLPoland	99.2%	Busy
RTX 4090 24 GB880 GB/s	10 vCPU · 23 GB RAM207 GB · ↓ 779	USUnited States	99.6%	Busy
RTX 4090 24 GB879 GB/s	10 vCPU · 23 GB RAM240 GB · ↓ 786	USUnited States	99.6%	Busy
RTX 4090 24 GB948 GB/s	32 vCPU · 93 GB RAM1,480 GB · ↓ 1,463	BRBrazil	95.7%	Busy
2×RTX 4090 24 GB880 GB/s	32 vCPU · 63 GB RAM1,609 GB · ↓ 627	SESweden	99.0%	Busy
2×RTX 4090 24 GB878 GB/s	32 vCPU · 63 GB RAM315 GB · ↓ 260	USUnited States	99.4%	Busy
RTX 4090 24 GB879 GB/s	28 vCPU · 63 GB RAM638 GB · ↓ 792	USUnited States	99.5%	Busy
2×RTX 4090 24 GB880 GB/s	32 vCPU · 62 GB RAM1,106 GB · ↓ 466	TRTürkiye	99.8%	Busy
RTX 4090 24 GB834 GB/s	8 vCPU · 31 GB RAM155 GB · ↓ 874	FRFrance	99.8%	Busy
2×RTX 4090 24 GB877 GB/s	32 vCPU · 62 GB RAM1,411 GB · ↓ 277	TRTürkiye	99.7%	Busy
2×RTX 4090 24 GB880 GB/s	32 vCPU · 63 GB RAM1,609 GB · ↓ 627	SESweden	99.0%	Busy
2×RTX 4090 24 GB877 GB/s	24 vCPU · 94 GB RAM636 GB · ↓ 927	FRFrance	99.1%	Busy
2×RTX 4090 24 GB879 GB/s	16 vCPU · 126 GB RAM231 GB · ↓ 840	PLPoland	98.9%	Busy
2×RTX 4090 24 GB879 GB/s	16 vCPU · 126 GB RAM226 GB · ↓ 761	PLPoland	98.9%	Busy
RTX 4090 24 GB877 GB/s	18 vCPU · 36 GB RAM410 GB · ↓ 689	USUnited States	99.7%	Busy
RTX 4090 24 GB877 GB/s	18 vCPU · 36 GB RAM398 GB · ↓ 689	USUnited States	99.7%	Busy
RTX 4090 24 GB877 GB/s	32 vCPU · 126 GB RAM671 GB · ↓ 786	SESweden	99.5%	Busy
2×RTX 4090 24 GB878 GB/s	13 vCPU · 75 GB RAM715 GB · ↓ 1,927	PLPoland	98.5%	Busy

RTX 4090
24 GBMalaysia
Busy
24 vCPU · 118 RAM · 394 disk · net ↓ 471
68₽per GPU/h1 GPU total · 67.9 ₽/h
Busy
93.3% · unverified
RTX 4090
24 GBIndia
Busy
20 vCPU · 125 RAM · 608 disk · net ↓ 453
68₽per GPU/h1 GPU total · 67.9 ₽/h
Busy
99.7% · verified
RTX 4090
24 GBPoland
Busy
24 vCPU · 47 RAM · 295 disk · net ↓ 1,183
65₽per GPU/h1 GPU total · 64.9 ₽/h
Busy
99.2% · verified
RTX 4090
24 GBUnited States
Busy
10 vCPU · 23 RAM · 207 disk · net ↓ 779
69₽per GPU/h1 GPU total · 68.9 ₽/h
Busy
99.6% · verified
RTX 4090
24 GBUnited States
Busy
10 vCPU · 23 RAM · 240 disk · net ↓ 786
69₽per GPU/h1 GPU total · 68.9 ₽/h
Busy
99.6% · verified
RTX 4090
24 GBBrazil
Busy
32 vCPU · 93 RAM · 1,480 disk · net ↓ 1,463
72₽per GPU/h1 GPU total · 71.9 ₽/h
Busy
95.7% · verified
2×RTX 4090
24 GBSweden
Busy
32 vCPU · 63 RAM · 1,609 disk · net ↓ 627
67₽per GPU/h2 GPU total · 134.9 ₽/h
Busy
99.0% · verified
2×RTX 4090
24 GBUnited States
Busy
32 vCPU · 63 RAM · 315 disk · net ↓ 260
62₽per GPU/h2 GPU total · 124.9 ₽/h
Busy
99.4% · verified
RTX 4090
24 GBUnited States
Busy
28 vCPU · 63 RAM · 638 disk · net ↓ 792
64₽per GPU/h1 GPU total · 63.9 ₽/h
Busy
99.5% · verified
2×RTX 4090
24 GBTürkiye
Busy
32 vCPU · 62 RAM · 1,106 disk · net ↓ 466
69₽per GPU/h2 GPU total · 137.9 ₽/h
Busy
99.8% · verified
RTX 4090
24 GBFrance
Busy
8 vCPU · 31 RAM · 155 disk · net ↓ 874
68₽per GPU/h1 GPU total · 67.9 ₽/h
Busy
99.8% · verified
2×RTX 4090
24 GBTürkiye
Busy
32 vCPU · 62 RAM · 1,411 disk · net ↓ 277
69₽per GPU/h2 GPU total · 137.9 ₽/h
Busy
99.7% · verified
2×RTX 4090
24 GBSweden
Busy
32 vCPU · 63 RAM · 1,609 disk · net ↓ 627
67₽per GPU/h2 GPU total · 134.9 ₽/h
Busy
99.0% · verified
2×RTX 4090
24 GBFrance
Busy
24 vCPU · 94 RAM · 636 disk · net ↓ 927
69₽per GPU/h2 GPU total · 137.9 ₽/h
Busy
99.1% · verified
2×RTX 4090
24 GBPoland
Busy
16 vCPU · 126 RAM · 231 disk · net ↓ 840
69₽per GPU/h2 GPU total · 137.9 ₽/h
Busy
98.9% · verified
2×RTX 4090
24 GBPoland
Busy
16 vCPU · 126 RAM · 226 disk · net ↓ 761
69₽per GPU/h2 GPU total · 137.9 ₽/h
Busy
98.9% · verified
RTX 4090
24 GBUnited States
Busy
18 vCPU · 36 RAM · 410 disk · net ↓ 689
69₽per GPU/h1 GPU total · 68.9 ₽/h
Busy
99.7% · verified
RTX 4090
24 GBUnited States
Busy
18 vCPU · 36 RAM · 398 disk · net ↓ 689
69₽per GPU/h1 GPU total · 68.9 ₽/h
Busy
99.7% · verified
RTX 4090
24 GBSweden
Busy
32 vCPU · 126 RAM · 671 disk · net ↓ 786
69₽per GPU/h1 GPU total · 68.9 ₽/h
Busy
99.5% · verified
2×RTX 4090
24 GBPoland
Busy
13 vCPU · 75 RAM · 715 disk · net ↓ 1,927
67₽per GPU/h2 GPU total · 133.9 ₽/h
Busy
98.5% · verified

Showing 561–580 of 1,877

PreviousPage 29 of 94Next

Questions

How much does a RTX 4090 24 GB rental cost?: From about 25 ₽ per GPU-hour at the lowest current offer. Prices change with availability, so the live figure above is the one to trust.
How is billing calculated?: Per minute while the server runs, from your balance. The price is fixed at launch. A server that fails to start is not billed.
Do I pay for the GPU when the server is stopped?: No GPU time is billed while the server is stopped. The disk is still billed, so keep only the data you need on it.
Is my data kept?: Data on the disk stays while the server exists. Deleting the server deletes its disk too, so download results first.
Which template should I start with?: Ollama for chatting with language models, ComfyUI or Forge for images and video, PyTorch with Jupyter for training and experiments, Ubuntu with CUDA when you want to install everything yourself.

Related models

RTX 5090 32 GB64₽
RTX 5080 16 GB38₽
RTX 5070 Ti 16 GB23₽
RTX 5070 12 GB27₽
RTX 5060 Ti 16 GB19₽
RTX 5060 Ti 8 GB19₽

All GPUs

Specifications

Video memory: 24 GB
Memory type: GDDR6X
Architecture: Ada Lovelace
Price from: 25₽/h

What it is good for

LLM (int8 / int4)

8Bfp16 · Fits
14Bint8 · Fits
32Bint4 · Fits
70BNot enough memory

What it is good for

Stable Diffusion 1.5Fits
SDXL: generationFits
FLUX.1 dev in fp8Fits
FLUX.1 dev in bf16Fits
Video: Wan 2.1 1.3BFits
Video: Wan 14B, HunyuanVideofrom 48 GB
SDXL LoRA trainingFits
QLoRA for 7B (4-bit)Fits
LoRA for 7B (fp16)Fits
Full fine-tuning of 7Bfrom 112 GB

Estimates for short context and typical settings. See the guide on memory for LLMs. How much VRAM an LLM needs: a table and a rule of thumb

Rental price

On-demand	per hour	per day
1 GPU	25₽	598₽
2 GPU	111₽	2,662₽
4 GPU	229₽	5,494₽
8 GPU	528₽	12,670₽

Interruptible: 15₽/h per GPUInterruptible bid price per GPU. The server can be stopped if someone outbids you.

Prices and availability come from the live catalog and update automatically.

Where it is available now

США106
Тайвань45
Канада26
Великобритания21
Норвегия20
Болгария17
Испания12
Япония11
Китай10
Южная Корея9
Германия8
Россия8
Эстония7
Венгрия6
Нидерланды6
Гонконг6
Румыния6
Финляндия6
ЮАР6
Бразилия5
Исландия5
Дания4
Франция4
Швеция4
Украина4
Малайзия4
Ирландия4
Польша3
Словакия3
Турция3
Швейцария2
Вьетнам2
Чили2
Северная Македония2
Австралия2
Чехия1
Италия1
Таиланд1
Сингапур1
Бельгия1
Беларусь1
Хорватия1

Countries: 42

How to start

Pick a server with the memory you need and a price you like in the catalog. Memory and type filters are on the catalog page.
Launch it with a template: Ollama for chatting with LLMs, ComfyUI or Forge for images and video, PyTorch + Jupyter for experiments.
Work in the browser or over SSH. Stop the server when you are done: GPU minutes stop being billed.

Pricing

Solutions for this GPU

Run LLMs on a rented GPU: Llama, Qwen, DeepSeek, Mistral
Stable Diffusion and FLUX on a rented GPU: Forge and SDXL online
ComfyUI online on your own server: FLUX, Wan and HunyuanVideo
Fine-tuning and LoRA training on a rented GPU
JupyterLab with a GPU for rent: PyTorch, CUDA and data science
GPU rendering on a rented server: Blender Cycles and 3D by the minute
Video generation on a rented GPU: Wan and HunyuanVideo in ComfyUI

Filters

Available onlyBusy servers are shown greyed and cannot be rented

GPU count

VRAM per GPU

Price, ₽/h per GPU

full range

4–9059

Region

USUnited States457
NONorway179
OMOman139
CACanada119
TWTaiwan113
JPJapan87

Reliability

Verified hosts only

More specs

CPU cores, minRAM, GB, minDisk, GB, minDownload, Mbit/s, minCUDA, minTFLOPS, min

Found: 1,877 serversAvailable: 396

Sort

GPU	Server	Region	Reliab.	Rent
RTX 4090 24 GB876 GB/s	24 vCPU · 118 GB RAM394 GB · ↓ 471	MYMalaysia	93.3%	Busy
RTX 4090 24 GB879 GB/s	20 vCPU · 125 GB RAM608 GB · ↓ 453	INIndia	99.7%	Busy
RTX 4090 24 GB899 GB/s	24 vCPU · 47 GB RAM295 GB · ↓ 1,183	PLPoland	99.2%	Busy
RTX 4090 24 GB880 GB/s	10 vCPU · 23 GB RAM207 GB · ↓ 779	USUnited States	99.6%	Busy
RTX 4090 24 GB879 GB/s	10 vCPU · 23 GB RAM240 GB · ↓ 786	USUnited States	99.6%	Busy
RTX 4090 24 GB948 GB/s	32 vCPU · 93 GB RAM1,480 GB · ↓ 1,463	BRBrazil	95.7%	Busy
2×RTX 4090 24 GB880 GB/s	32 vCPU · 63 GB RAM1,609 GB · ↓ 627	SESweden	99.0%	Busy
2×RTX 4090 24 GB878 GB/s	32 vCPU · 63 GB RAM315 GB · ↓ 260	USUnited States	99.4%	Busy
RTX 4090 24 GB879 GB/s	28 vCPU · 63 GB RAM638 GB · ↓ 792	USUnited States	99.5%	Busy
2×RTX 4090 24 GB880 GB/s	32 vCPU · 62 GB RAM1,106 GB · ↓ 466	TRTürkiye	99.8%	Busy
RTX 4090 24 GB834 GB/s	8 vCPU · 31 GB RAM155 GB · ↓ 874	FRFrance	99.8%	Busy
2×RTX 4090 24 GB877 GB/s	32 vCPU · 62 GB RAM1,411 GB · ↓ 277	TRTürkiye	99.7%	Busy
2×RTX 4090 24 GB880 GB/s	32 vCPU · 63 GB RAM1,609 GB · ↓ 627	SESweden	99.0%	Busy
2×RTX 4090 24 GB877 GB/s	24 vCPU · 94 GB RAM636 GB · ↓ 927	FRFrance	99.1%	Busy
2×RTX 4090 24 GB879 GB/s	16 vCPU · 126 GB RAM231 GB · ↓ 840	PLPoland	98.9%	Busy
2×RTX 4090 24 GB879 GB/s	16 vCPU · 126 GB RAM226 GB · ↓ 761	PLPoland	98.9%	Busy
RTX 4090 24 GB877 GB/s	18 vCPU · 36 GB RAM410 GB · ↓ 689	USUnited States	99.7%	Busy
RTX 4090 24 GB877 GB/s	18 vCPU · 36 GB RAM398 GB · ↓ 689	USUnited States	99.7%	Busy
RTX 4090 24 GB877 GB/s	32 vCPU · 126 GB RAM671 GB · ↓ 786	SESweden	99.5%	Busy
2×RTX 4090 24 GB878 GB/s	13 vCPU · 75 GB RAM715 GB · ↓ 1,927	PLPoland	98.5%	Busy

RTX 4090
24 GBMalaysia
Busy
24 vCPU · 118 RAM · 394 disk · net ↓ 471
68₽per GPU/h1 GPU total · 67.9 ₽/h
Busy
93.3% · unverified
RTX 4090
24 GBIndia
Busy
20 vCPU · 125 RAM · 608 disk · net ↓ 453
68₽per GPU/h1 GPU total · 67.9 ₽/h
Busy
99.7% · verified
RTX 4090
24 GBPoland
Busy
24 vCPU · 47 RAM · 295 disk · net ↓ 1,183
65₽per GPU/h1 GPU total · 64.9 ₽/h
Busy
99.2% · verified
RTX 4090
24 GBUnited States
Busy
10 vCPU · 23 RAM · 207 disk · net ↓ 779
69₽per GPU/h1 GPU total · 68.9 ₽/h
Busy
99.6% · verified
RTX 4090
24 GBUnited States
Busy
10 vCPU · 23 RAM · 240 disk · net ↓ 786
69₽per GPU/h1 GPU total · 68.9 ₽/h
Busy
99.6% · verified
RTX 4090
24 GBBrazil
Busy
32 vCPU · 93 RAM · 1,480 disk · net ↓ 1,463
72₽per GPU/h1 GPU total · 71.9 ₽/h
Busy
95.7% · verified
2×RTX 4090
24 GBSweden
Busy
32 vCPU · 63 RAM · 1,609 disk · net ↓ 627
67₽per GPU/h2 GPU total · 134.9 ₽/h
Busy
99.0% · verified
2×RTX 4090
24 GBUnited States
Busy
32 vCPU · 63 RAM · 315 disk · net ↓ 260
62₽per GPU/h2 GPU total · 124.9 ₽/h
Busy
99.4% · verified
RTX 4090
24 GBUnited States
Busy
28 vCPU · 63 RAM · 638 disk · net ↓ 792
64₽per GPU/h1 GPU total · 63.9 ₽/h
Busy
99.5% · verified
2×RTX 4090
24 GBTürkiye
Busy
32 vCPU · 62 RAM · 1,106 disk · net ↓ 466
69₽per GPU/h2 GPU total · 137.9 ₽/h
Busy
99.8% · verified
RTX 4090
24 GBFrance
Busy
8 vCPU · 31 RAM · 155 disk · net ↓ 874
68₽per GPU/h1 GPU total · 67.9 ₽/h
Busy
99.8% · verified
2×RTX 4090
24 GBTürkiye
Busy
32 vCPU · 62 RAM · 1,411 disk · net ↓ 277
69₽per GPU/h2 GPU total · 137.9 ₽/h
Busy
99.7% · verified
2×RTX 4090
24 GBSweden
Busy
32 vCPU · 63 RAM · 1,609 disk · net ↓ 627
67₽per GPU/h2 GPU total · 134.9 ₽/h
Busy
99.0% · verified
2×RTX 4090
24 GBFrance
Busy
24 vCPU · 94 RAM · 636 disk · net ↓ 927
69₽per GPU/h2 GPU total · 137.9 ₽/h
Busy
99.1% · verified
2×RTX 4090
24 GBPoland
Busy
16 vCPU · 126 RAM · 231 disk · net ↓ 840
69₽per GPU/h2 GPU total · 137.9 ₽/h
Busy
98.9% · verified
2×RTX 4090
24 GBPoland
Busy
16 vCPU · 126 RAM · 226 disk · net ↓ 761
69₽per GPU/h2 GPU total · 137.9 ₽/h
Busy
98.9% · verified
RTX 4090
24 GBUnited States
Busy
18 vCPU · 36 RAM · 410 disk · net ↓ 689
69₽per GPU/h1 GPU total · 68.9 ₽/h
Busy
99.7% · verified
RTX 4090
24 GBUnited States
Busy
18 vCPU · 36 RAM · 398 disk · net ↓ 689
69₽per GPU/h1 GPU total · 68.9 ₽/h
Busy
99.7% · verified
RTX 4090
24 GBSweden
Busy
32 vCPU · 126 RAM · 671 disk · net ↓ 786
69₽per GPU/h1 GPU total · 68.9 ₽/h
Busy
99.5% · verified
2×RTX 4090
24 GBPoland
Busy
13 vCPU · 75 RAM · 715 disk · net ↓ 1,927
67₽per GPU/h2 GPU total · 133.9 ₽/h
Busy
98.5% · verified

Showing 561–580 of 1,877

PreviousPage 29 of 94Next

США106
Тайвань45
Канада26
Великобритания21
Норвегия20
Болгария17
Испания12
Япония11
Китай10
Южная Корея9
Германия8
Россия8
Эстония7
Венгрия6
Нидерланды6
Гонконг6
Румыния6
Финляндия6
ЮАР6
Бразилия5
Исландия5
Дания4
Франция4
Швеция4
Украина4
Малайзия4
Ирландия4
Польша3
Словакия3
Турция3
Швейцария2
Вьетнам2
Чили2
Северная Македония2
Австралия2
Чехия1
Италия1
Таиланд1
Сингапур1
Бельгия1
Беларусь1
Хорватия1

Countries: 42

Questions

How much does a RTX 4090 24 GB rental cost?: From about 25 ₽ per GPU-hour at the lowest current offer. Prices change with availability, so the live figure above is the one to trust.
How is billing calculated?: Per minute while the server runs, from your balance. The price is fixed at launch. A server that fails to start is not billed.
Do I pay for the GPU when the server is stopped?: No GPU time is billed while the server is stopped. The disk is still billed, so keep only the data you need on it.
Is my data kept?: Data on the disk stays while the server exists. Deleting the server deletes its disk too, so download results first.
Which template should I start with?: Ollama for chatting with language models, ComfyUI or Forge for images and video, PyTorch with Jupyter for training and experiments, Ubuntu with CUDA when you want to install everything yourself.

Related models

RTX 5090 32 GB64₽
RTX 5080 16 GB38₽
RTX 5070 Ti 16 GB23₽
RTX 5070 12 GB27₽
RTX 5060 Ti 16 GB19₽
RTX 5060 Ti 8 GB19₽

All GPUs