Claude API pricing
Claude API usage is billed per token, with input and output tokens priced separately and output usually costing more. Your bill is therefore driven by how many tokens you send, how many the model returns, and which model you chose — not by the number of requests alone.
What determines the cost of a request?
Three things, in this order: the model, the number of input tokens, and the number of output tokens. Prices are quoted per million tokens, and input and output have separate rates, so a short prompt with a long answer can cost more than a long prompt with a short one.
- Input tokens — everything you send, including system prompts and context
- Output tokens — everything the model returns, normally the more expensive side
- Model choice — rates differ by an order of magnitude across a family
How to estimate Claude API spend
Multiply your average input tokens by the input rate and your average output tokens by the output rate, then multiply by the number of requests. The arithmetic is simple; the hard part is knowing your real token counts, which is why per-request visibility matters more than a forecast.
Using Claude models through Astrodyne
Astrodyne is an OpenAI-compatible gateway: you name a model in the request, point your client at the Astrodyne base URL and use an Astrodyne key. Billing is one prepaid balance rather than a separate account and invoice per provider, and every request records its own tokens and charge. Which models your account can call is shown in your console and returned by GET /v1/models — that list is the authoritative answer, and access is invite-only while we roll out.
Why can Astrodyne cost less?
Astrodyne aggregates demand and maintains the provider relationships behind one API. That lets eligible accounts access commercial terms normally associated with larger buyers, without negotiating each provider relationship themselves. Availability and pricing advantages vary by model.
Frequently asked questions
How is Claude API pricing calculated?
Per token, with input and output priced separately and quoted per million tokens. Multiply your input tokens by the input rate, your output tokens by the output rate, and add them.
How can I reduce Claude API costs?
Send fewer input tokens by trimming context, cap output length where you can, and use a smaller model for the requests that do not need a larger one. Beyond that, per-request cost visibility tells you which calls are actually expensive, which is usually a different set from the ones you would guess.
Is Astrodyne cheaper than using Claude directly?
Only where Astrodyne can show it. Astrodyne publishes a price comparison for a model when it holds both an approved Astrodyne price and a current published direct rate to compare against. Where that comparison is not available, the pricing page says so rather than implying a saving. Availability and any pricing advantage vary by model.
Which Claude models can I use through Astrodyne?
Model availability is account-specific. The models your key can call are listed in your console and returned by GET /v1/models — that list is the authoritative answer for your account.
Do I have to change my code?
Yes. Point an OpenAI-compatible client at the Astrodyne base URL with an Astrodyne API key. Streaming, tool calls and the response shape are unchanged, so the migration is a base URL and a key.