Amazon Bedrock is a service offered by AWS that allows users to access and utilize foundation models from various AI providers. Below is a pricing overview for different models available through Amazon Bedrock, as summarized by Future Cloud:
1. AI21 Labs Models:
- Jamba 1.5 Large: $0.002 per 1,000 input tokens, $0.008 per 1,000 output tokens.
- Jamba 1.5 Mini: $0.0002 per 1,000 input tokens, $0.0004 per 1,000 output tokens.
- Jurassic-2 Mid: $0.0125 per 1,000 input tokens, $0.0125 per 1,000 output tokens.
- Jurassic-2 Ultra: $0.0188 per 1,000 input tokens, $0.0188 per 1,000 output tokens.
2. Anthropic Models (East and West U.S. regions):
- Claude 3.5 Sonnet: $0.003 per 1,000 input tokens, $0.015 per 1,000 output tokens; batch inference pricing is $0.0015 per 1,000 input tokens, $0.0075 per 1,000 output tokens.
- Claude 3.5 Haiku: $0.001 per 1,000 input tokens, $0.005 per 1,000 output tokens; batch inference pricing is $0.0005 per 1,000 input tokens, $0.0025 per 1,000 output tokens.
3. Cohere Models:
- Command: $0.0015 per 1,000 input tokens, $0.0020 per 1,000 output tokens.
- Command-Light: $0.0003 per 1,000 input tokens, $0.0006 per 1,000 output tokens.
4. Meta Llama Models:
- Llama 2 Chat (13B): $0.00075 per 1,000 input tokens, $0.001 per 1,000 output tokens.
5. Mistral AI Models:
- Mistral 7B Instruct: $0.00015 per 1,000 input tokens, $0.0002 per 1,000 output tokens.
- Mistral Large: $0.008 per 1,000 input tokens, $0.024 per 1,000 output tokens.
6. Stability AI Models (Multimodal Models):
- Stable Diffusion XL (v1.0): $0.04 per image for standard quality, $0.08 per image for premium quality.
These prices provide an overview of on-demand and batch inference pricing for various models available through Amazon Bedrock. Users can choose the appropriate model and pricing plan based on their needs. Please note that these prices may vary depending on specific usage and AWS’s pricing strategy.