Cookies on AIGROW

We use necessary cookies to run the site. With your permission, we also use cookies for analytics, session recordings and marketing attribution.

Cookie Policy

Claude API

Cheap Claude API: Opus 5.5, Fable 5.2 and Sonnet 5.5 below list

AIGROW sells prepaid credit for one API endpoint that serves Claude Opus 5.5, Claude Fable 5.2 and Claude Sonnet 5.5. Credit costs a quarter of its face value, so Opus 5.5 comes to $1.60 per million input tokens and $8 per million output tokens. Anthropic’s list price for the same model is $4 and $20. Sonnet 5.5 costs $2 and $10 at list, so the larger model costs less here.

See every model and price

List prices checked 12 November 2026. The AIGROW rates below are read from the same file the endpoint bills with.

Claude API prices, list and through AIGROW credit

USD per million tokens: input / output, with prompt-cache reads and writes beneath.

ModelWith AIGROW creditAnthropic list price
Claude Opus 5.5claude-opus-5.5$1.60 / $8.00$0.20 cache read$3.20 cache write$4.00 / $20.00$0.20 cache read$8.00 cache write
Claude Fable 5.2claude-fable-5.2$3.20 / $16.00$0.20 cache read$6.40 cache write$8.00 / $40.00$0.20 cache read$16.00 cache write
Claude Fable 5.1claude-fable-5.1$4.00 / $20.00$0.25 cache read$8.00 cache write$10.00 / $50.00$0.25 cache read$20.00 cache write
Claude Sonnet 5.5claude-sonnet-5.5$0.80 / $4.00$0.075 cache read$1.60 cache write$2.00 / $10.00$0.10 cache read$4.00 cache write

How we get these numbers: each model has a rate in credit. You pay a quarter of it, because $25 buys $100 of credit. Every cache write on these models goes to the 1-hour cache and costs 2 times the input rate, so the list price beside it is Anthropic’s for a 1-hour write. List prices are Anthropic’s, from its pricing page.

Opus 5.5 for less than Sonnet 5.5’s list price

For input and output, Opus 5.5 on AIGROW credit costs $1.60 and $8 per million tokens. Sonnet 5.5 at Anthropic’s list price costs $2 and $10. If you use Sonnet to keep costs down, Opus here costs 20% less per token.

Cache reads are the exception. On Opus 5.5 they cost the list price, $0.20 per million, and on Sonnet 5.5 here they cost $0.075. On a workload that reads most of each request from the prompt cache, such as a coding agent, Opus here costs more than Sonnet at list: 250 requests a working day come to $138.60 a month on Opus 5.5 through credit and $126 on Sonnet 5.5 at list. Work out a month of Claude Code on any model.

What a month costs

Example: 20 million input tokens and 4 million output tokens in a month, no caching.

A month of Claude API use through AIGROW credit and at list
ModelThrough AIGROW creditAt listYou save
Claude Opus 5.5$64$160$96
Claude Fable 5.2$128$320$192
Claude Fable 5.1$160$400$240
Claude Sonnet 5.5$32$80$48

A $25 pack gives $100 of credit: enough for 15 million Opus 5.5 input tokens or 3 million output tokens. Monthly plans add 10% to 25% more credit than the pack at the same price.

Other ways to pay less for Claude

  • Prompt caching. Repeated context read from the cache costs a fraction of the input price on every Claude model. It works the same way on this endpoint.
  • Pick the model per task. Sonnet 5.5 here costs 60% less than its list price on input and output; Opus 5.5 here costs less per input and output token than Sonnet at list.

Connect in two settings

A base URL and a key that starts with ag_, created in API access on the dashboard.

  • Anthropic SDK and Claude Code: base URL https://aigrow.gg
  • OpenAI SDK, Cline, Aider, OpenCode: base URL https://aigrow.gg/v1
  • Model id: as in the table, for example claude-opus-5.5. Anthropic’s own spelling (claude-opus-5-5) and the router spelling (anthropic/claude-opus-5.5) work too.

Anthropic SDK (Python)

Set base_url and api_key when you create the client.

from anthropic import Anthropic

client = Anthropic(
    base_url="https://aigrow.gg",
    api_key="ag_your_key_here",
)

msg = client.messages.create(
    model="claude-opus-5.5",
    max_tokens=512,
    messages=[{"role": "user", "content": "Hello"}],
)
print(msg.content[0].text)

Claude Code

Set these environment variables, then run it as usual. The last one keeps its background calls on a model the endpoint serves.

export ANTHROPIC_BASE_URL="https://aigrow.gg"
export ANTHROPIC_AUTH_TOKEN="ag_your_key_here"
export ANTHROPIC_MODEL="claude-opus-5.5"
export ANTHROPIC_DEFAULT_HAIKU_MODEL="claude-sonnet-5.5"

claude

The setup guide covers every other tool, and the Claude Code page walks through its settings one by one.

What is billed and what is not

A request for a model outside the catalogue is refused before it is sent, and nothing is billed. A 429 (slow down) and a 502 or 503 (temporary error) are not billed either. When your balance reaches zero, requests return 402 until you redeem more credit.

Claude API questions

Yes. AIGROW credit costs a quarter of its face value, so Claude Opus 5.5 costs $1.60 per million input tokens and $8 per million output tokens, against Anthropic’s list price of $4 and $20. Prompt caching lowers the cost further on repeated context.

$1.60 for input, $8 for output, $0.20 for a cache read and $3.20 for a cache write, which goes to the 1-hour cache. The context window is 1,000,000 tokens.

Per input and output token, Opus 5.5 here costs 20% less than Sonnet 5.5 at its list price. Cache reads on Opus 5.5 cost the list price, so on cache-heavy work such as a coding agent Opus here costs more: a month of 250 coding agent requests a working day costs $138.60 on Opus 5.5 through credit and $126 on Sonnet 5.5 at list.

Both. The Anthropic SDK takes the base URL https://aigrow.gg and the OpenAI SDK takes https://aigrow.gg/v1, with the same ag_ key and a Claude model id such as claude-opus-5.5.

Yes. Set ANTHROPIC_BASE_URL to https://aigrow.gg, ANTHROPIC_AUTH_TOKEN to your key, and ANTHROPIC_DEFAULT_HAIKU_MODEL to claude-sonnet-5.5 so background calls run on a model the endpoint serves.

Only if you choose one. A pack is a one-off payment. A monthly plan charges the same amount each month, adds 10% to 25% more credit and can be cancelled at any time.

No. AIGROW is independent and runs its own endpoint. Model names identify what the endpoint serves and do not imply a partnership or endorsement.

Claude Opus 5.5 at $1.60 / $8 per million tokens.

Start with the $25 pack, run your real workload against it, and check the dashboard before buying more.

Buy API creditSee every model and price

Monthly plans or one-off packs · redemption code delivered at once · usage dashboard included