How much does a Claude API key cost?

The key is free. You pay per token used: see the table above for the current Claude input, output, cache-write and cache-read rates per million tokens.

Which is cheaper, Claude or GPT or Gemini?

It depends on the model tier you pick, not the vendor: small/fast models on every provider land in the same order of magnitude, while flagship models cost several times more per token. Compare the rows above for the specific models you plan to call.

What is an AI gateway and does it add cost?

A gateway puts many providers behind one OpenAI-compatible endpoint and one key. CodeGateway adds a transparent flat markup (currently 1.07×) instead of a subscription fee.

How are cache writes and cache reads billed?

Cache writes cost slightly more than normal input tokens because the provider stores the context; cache reads cost a fraction of input tokens. Reusing a long system prompt is usually cheaper with caching enabled.

Do I need a monthly subscription?

No. Top up a balance and each call deducts the exact token cost (precision to $10⁻¹⁰). The balance never expires, there is no monthly minimum, and top-ups start at $5.

An API key itself costs nothing — on every major provider you pay per token you actually send and receive, not for the key.
List prices are quoted per million tokens (MTok). Output tokens usually cost 3–5× input tokens, and cache reads are the cheapest way to re-send a long prompt.
CodeGateway bills the provider list price times a flat markup (currently 1.07×). No subscription or monthly minimum; top up from $5, and your balance never expires.

Claude vs GPT vs Gemini: API price per million tokens

Live rates for every model CodeGateway proxies. Input, output, cache write and cache read, in USD per million tokens.

Model	Input /MTok	Output /MTok	Cache Write /MTok	Cache Read /MTok	Unit
claude-sonnet-4-5Anthropic	$3.2100	$16.0500	$4.0125	$0.3210	/MTok
claude-sonnet-4-5-20250929Anthropic	$3.2100	$16.0500	$4.0125

API pricing FAQ

How much does a Claude API key cost?: The key is free. You pay per token used: see the table above for the current Claude input, output, cache-write and cache-read rates per million tokens.
Which is cheaper, Claude or GPT or Gemini?: It depends on the model tier you pick, not the vendor: small/fast models on every provider land in the same order of magnitude, while flagship models cost several times more per token. Compare the rows above for the specific models you plan to call.
What is an AI gateway and does it add cost?: A gateway puts many providers behind one OpenAI-compatible endpoint and one key. CodeGateway adds a transparent flat markup (currently 1.07×) instead of a subscription fee.
How are cache writes and cache reads billed?: Cache writes cost slightly more than normal input tokens because the provider stores the context; cache reads cost a fraction of input tokens. Reusing a long system prompt is usually cheaper with caching enabled.
Do I need a monthly subscription?: No. Top up a balance and each call deducts the exact token cost (precision to $10⁻¹⁰). The balance never expires, there is no monthly minimum, and top-ups start at $5.

One key, every model — start free

Create a key, send one request, and see the exact per-token cost in your usage log.

Create your API key