Supply Claude API capacity
Which Claude model can you serve, through which channel, and at what rate? We review offers for usage you are authorized to supply through the Claude API or Amazon Bedrock.

Claude API / Amazon Bedrock
Claude
A monthly offer, worked through
Illustrative quantities and prices. Replace the model, limits and commercial terms with your own.
- Model and channel
- Claude Sonnet via the direct API. Include the exact model ID and any residency commitment.
- Available throughput
- 30 RPM, 60,000 input TPM and 12,000 output TPM reserved after internal use.
- Monthly delivery
- Up to 20 million input tokens and 4 million output tokens over one month.
- Asking price
- USD 2 per million input tokens and USD 8 per million output tokens. At the stated volumes, USD 72 in total.
What to include for each channel
Direct Claude API
Give the exact model ID and the limits that apply to your organization or workspace. Identify the part of each limit you can reserve for this offer.
- Model ID, service region or residency constraint, and delivery dates.
- Requests per minute (RPM), input tokens per minute and output tokens per minute. Mark any limit shared across customers or models.
- Maximum billable usage per month, currency, and separate rates per million input and output tokens. Include cache pricing if applicable.
Claude on Amazon Bedrock
Use the Bedrock model ID or inference profile ID from your planned deployment. Specify the AWS Region where requests enter and any permitted routing regions.
- On-demand or provisioned capacity, model or profile ID, and any regional restrictions.
- Applicable quota units, shared traffic, and throughput actually available during the quoted hours. Some quotas count both input and output tokens.
- Delivery method allowed by your agreement, unit price, and any minimum commitment or provisioned-capacity fee.
Before you send the offer
Which permission should I confirm?
Check that your provider agreement or written authorization covers the proposed delivery to a third party. Describe the arrangement and any limits. A console quota alone does not establish that right. Keep keys and sign-in credentials private.
Is a high TPM limit enough for a monthly commitment?
TPM is a rate. Monthly availability also depends on the hours you can serve, the billable volume you can commit and other traffic sharing the limit. Give each figure separately.
Can I quote both delivery channels?
Yes. Use a separate line per channel with its model, region, capacity and price. If both offers draw on the same allocation, state the combined amount you can commit.
Preparing an API capacity quote
How to quote AI API capacity
Separate throughput from billable usage, spell out available models and limits, and quote the unit price for each kind of token.