AI costs
AI is an extra service that can be enabled and disabled per camera at any time. If enabled, it’s typically few dollars/month per camera and depends on the amount of motion and AI services as described below.
total AI cost = [number of (motion) triggers] x [AI cost per image]
Generative AI cost
The cost of Generative AI per image depends on the model and image resolution.
Resolution
Frames with a higher resolution consume GenAI tokens. Full high definition (FHD) consumes usually twice more tokens than high definition (HD) video.
Model
The model determines which LLM (Large Language Model) will be used to perform queries on. Some models are more sophisticated than others, hence the Advanced, Pro, and Lite naming convention. The specific names of these models are:
Name | Advanced | Pro | Lite |
|---|---|---|---|
LLM | Claude 3.5 Sonnet | Nova Pro | Nova Lite |
Use cases |
|
|
|
Models, resolution and costs
Below is approximate cost per 1,000 images
Model | HD | FHD |
Nova Lite | $0.10 | $0.20 |
Nova Pro | $0.75 | $1.50 |
Claude 3.5 | $3.00 | $6.00 |
More information can be found using the links below:
https://aws.amazon.com/bedrock/pricing/
https://docs.aws.amazon.com/bedrock/latest/userguide/models-supported.html
https://docs.aws.amazon.com/nova/latest/userguide/modalities-image.html
https://aws.amazon.com/ai/generative-ai/nova/understanding/
Facial Recognition and Object Recognition AI cost
The cost per 1,000 images is $1, with discounts available for large volumes.
https://aws.amazon.com/rekognition/pricing/