Product

En

Pricing

SiliconCloud

SiliconCloud

SiliconCloud

Fine-tuning

Model

Price

Multimodal

*SiliconCloud charges based on the total number of tokens in your fine-tuning dataset.

Fine-tuning

Fine-tuning

Model

Price

Model

Price

Multimodal

*SiliconCloud charges based on the total number of tokens in your fine-tuning dataset.

*SiliconCloud charges based on the total number of tokens in your fine-tuning dataset.

Deployment Service

Deployment Service

Serverless Deployment

Serverless Deployment

Model

Price

Base model

Fine-tuned model

price of corresponding base model*1.5

Model

Price

Base model

Fine-tuned model

price of corresponding base model*1.5

Multimodal

On-demand Deployment & Reserved Capacity

On-demand Deployment & Reserved Capacity

Prepaid Duration

Prepaid Duration

GPU Type

GPU Type*

Price

Price

Pay-as-you-go GPU instance

Compute Unit A

Compute Unit B

$0.35/h

Pay-as-you-go Inference Engine

Compute Unit A

Compute Unit B

$0.30/h

$0.23/h

$0.80/h

Reserved Capacity GPU instance

With Inference Engine

Compute Unit A

Compute Unit B

Product

Product Type*

Price

Pay-as-you-go GPU instance

Compute Unit A

Compute Unit B

Pay-as-you-go Inference Engine

Compute Unit A

Compute Unit B

$0.30/h

$0.80/h

Reserved Capacity GPU instance

with Inference Engine

Compute Unit A

Compute Unit B

Multimodal

*Compute Unit A is lower cost, Comoute Unit B is lower latency.

*Compute Unit A is lower cost, Comoute Unit B is lower latency.

Software Subscription
Dedicated email support.

Software

Price

OneDiff

/NVIDIA Hopper architecture GPUs

OneDiff

/Other GPUs

*The current product price is only limited to use outside mainland China.

Software Subscription

Dedicated email support

Software

Price

OneDiff

/NVIDIA Hopper architecture GPUs

OneDiff

/Other GPUs

*The current product price is only limited to use outside mainland China.

Deployment Service

Serverless Deployment

Model

Price

Base model

Fine-tuned model

price of corresponding base model*1.5

Multimodal

*Compute Unit A is lower cost, Comoute Unit B is lower latency.

On-demand Deployment

Prepaid Duration

GPU Type*

Price

Pay-as-you-go GPU instance

Compute Unit A

Compute Unit B

Pay-as-you-go Inference Engine

Compute Unit A

Compute Unit B

$0.30/h

$0.80/h

Reserved Capacity GPU instance

and Accleration Framework

Compute Unit A

Compute Unit B

Multimodal

Dedicated email support
Software Subscription

Software

Price

OneDiff

/NVIDIA Hopper architecture GPUs

OneDiff

/Other GPUs

*The current product price is only limited to use outside mainland China.

Frequently Asked Questions

Can I try SiliconCloud before subscribing?

Can I try OneDiff or SiliconLLM before subscribing?

Can I get a refund if I cancel my subscription before it expires?

Product

En

Product

En

Pricing

Pricing

Pricing

En
En
Frequently Asked Questions

Can I try SiliconCloud before subscribing?

Can I try OneDiff or SiliconLLM before subscribing?

Can I get a refund if I cancel my subscription before it expires?

Can I try SiliconCloud before subscribing?

Can I try OneDiff or SiliconLLM before subscribing?

Can I get a refund if I cancel my subscription before it expires?

Frequently Asked Questions

Can I try SiliconCloud before subscribing?

Can I try OneDiff or SiliconLLM before subscribing?

Can I get a refund if I cancel my subscription before it expires?

Can I try SiliconCloud before subscribing?

Can I try OneDiff or SiliconLLM before subscribing?

Can I get a refund if I cancel my subscription before it expires?