AI Model Endpoints
AI Model Endpoints offer stable APIs for text, images, video, and more through one unified interface.
🤖
View All >Chat
High-Performance GPUs Transparent Pricing
Per-second billing for maximum savings , Free 100 GB container storage
A100 SXM
$1.60/hour
($896/month)
Data center-grade performance for large-scale training
- 80GB VRAM
- Enterprise-Ready Utilization
- Powered by the NVIDIA Ampere Architecture
Most popular
RTX 4090
$0.42/hour
($252/month)
Ideal for high-performance inference and medium-scale training
- 24GB VRAM
- Fourth-Gen Tensor Cores
- RTX Video Super Resolution and NVIDIA Broadcast
H100 SXM
$2.56/hour
($1580/month)
Flagship compute power for next-generation AI workloads
- 80GB VRAM
- NVIDIA Hopper™ architecture
- Includes a dedicated Transformer Engine
Billed by the second,—pay only for what you use. view pricing list
Best Prices for 10,000+ GPUs
Best Prices for 10,000+ GPUs
A100 SXM
$1.60/hour
($896/month)
Data center-grade performance for large-scale training
- 80GB VRAM
- Enterprise-Ready Utilization
- Powered by the NVIDIA Ampere Architecture
Most popular
RTX 4090
$0.42/hour
($252/month)
Ideal for high-performance inference and medium-scale training
- 24GB VRAM
- Fourth-Gen Tensor Cores
- RTX Video Super Resolution and NVIDIA Broadcast
H100 SXM
$2.56/hour
($1580/month)
Flagship compute power for next-generation AI workloads
- 80GB VRAM
- NVIDIA Hopper™ architecture
- Includes a dedicated Transformer Engine
Billed by the second,—pay only for what you use. view pricing list
Best Prices for 10,000+ GPUs
Best Prices for 10,000+ GPUs
A100 SXM
$1.60/hour
($896/month)
Data center-grade performance for large-scale training
- 80GB VRAM
- Enterprise-Ready Utilization
- Powered by the NVIDIA Ampere Architecture
Most popular
RTX 4090
$0.42/hour
($252/month)
Ideal for high-performance inference and medium-scale training
- 24GB VRAM
- Fourth-Gen Tensor Cores
- RTX Video Super Resolution and NVIDIA Broadcast
H100 SXM
$2.56/hour
($1580/month)
Flagship compute power for next-generation AI workloads
- 80GB VRAM
- NVIDIA Hopper™ architecture
- Includes a dedicated Transformer Engine
Billed by the second,—pay only for what you use. view pricing list
System Images Community Built by global developers.
Upload your own image or directly use the community-verified high-quality solution.
System Images Community Built by global developers.
Upload your own image or directly use the community-verified high-quality solution.
System Images Community Built by global developers.
Upload your own image or directly use the community-verified high-quality solution.
System Images Community Built by global developers.
Upload your own image or directly use the community-verified high-quality solution.
Powerful GPU Cluster for AI Computing
24/7
support
99%
uptime
15min
reply
Supports creating virtual
machines or containers
Choose instance types on demand, which is flexible and elastic while also taking security and isolation into consideration.
Model images support
rapid deployment
The platform provides a variety of model images and built-in applications such as JupyterLab, which greatly reduces your deployment time.
Multiple storage
mode options
Offers two storage options: highly reliable, scalable SSD system disk, and object storage that supports shared access across multiple instances.
Flexible leasing methods and
extremely low price
Multiple billing methods, supports postpaid model, with costs based on actual usage time. Save up to 50% on model costs without compromising performance.
Powerful GPU Cluster for AI Computing
24/7
support
99%
uptime
15min
reply
Supports creating virtual
machines or containers
Choose instance types on demand, which is flexible and elastic while also taking security and isolation into consideration.
Model images support
rapid deployment
The platform provides a variety of model images and built-in applications such as JupyterLab, which greatly reduces your deployment time.
Multiple storage
mode options
Offers two storage options: highly reliable, scalable SSD system disk, and object storage that supports shared access across multiple instances.
Flexible leasing methods and
extremely low price
Multiple billing methods, supports postpaid model, with costs based on actual usage time. Save up to 50% on model costs without compromising performance.
Powerful GPU Cluster for AI Computing
24/7
support
99%
uptime
15min
reply
Supports creating virtual
machines or containers
Choose instance types on demand, which is flexible and elastic while also taking security and isolation into consideration.
Model images support
rapid deployment
The platform provides a variety of model images and built-in applications such as JupyterLab, which greatly reduces your deployment time.
Multiple storage
mode options
Offers two storage options: highly reliable, scalable SSD system disk, and object storage that supports shared access across multiple instances.
Flexible leasing methods and
extremely low price
Multiple billing methods, supports postpaid model, with costs based on actual usage time. Save up to 50% on model costs without compromising performance.
Powerful GPU Cluster for AI Computing
24/7
support
99%
uptime
15min
reply
Supports creating virtual
machines or containers
Choose instance types on demand, which is flexible and elastic while also taking security and isolation into consideration.
Model images support
rapid deployment
The platform provides a variety of model images and built-in applications such as JupyterLab, which greatly reduces your deployment time.
Multiple storage
mode options
Offers two storage options: highly reliable, scalable SSD system disk, and object storage that supports shared access across multiple instances.
Flexible leasing methods and
extremely low price
Multiple billing methods, supports postpaid model, with costs based on actual usage time. Save up to 50% on model costs without compromising performance.
Contact