Skip to content
1折AI
OpenAI / GPT and Image APIs are live

Claude API81% off list price

Every model and cache tier is billed at one discount, and every call is itemized in your usage log.

Minimum top-up ¥100. Unused credit is refundable to your original payment method within 7 days of top-up.

We acceptWeChat PayAlipayStripeUSDT
One Claude Code call
200 · 1840 ms
modelclaude-sonnet-5-5
Input
34,120
Cache read
94,310
Cache write
2,048
Output
4,215
List price$0.1249
Your priceSave 81%$0.0237
Sample calculation, not a real call. Your price = List price × 0.19

Claude, GPT, and Image APIs

Claude API
Live
Fable, Opus, Sonnet, and Haiku.

Sonnet 5.5 output, per 1M tokens

List price
$10.00x
Your price
$1.90x
GPT API
Live
OpenAI / GPT models are live. Connect with the OpenAI SDK or Codex CLI.
Image API
Live
gpt-image-1.5 and chatgpt-image-latest are available. See the model catalog for endpoints and pricing.

One discount, every model

Input, output, cache reads, and cache writes are all billed at list price times one discount.

Your price=List price×0.19

Some platforms bill with a top-up ratio and a multiplier. We use list price and one discount.

How our pricing works (in Chinese)

Claude pricing

Per 1M tokens, USD.

See all models
Per 1M tokens
  • claude-fable-5-1
    Input
    List $10.00$1.90
    Output
    List $50.00$9.50
    Cache read
    List $0.25$0.048
    Cache write
    List $12.50$2.38
  • claude-opus-5-5
    Input
    List $4.00$0.76
    Output
    List $20.00$3.80
    Cache read
    List $0.20$0.038
    Cache write
    List $5.00$0.95
  • claude-sonnet-5-5
    Input
    List $2.00$0.38
    Output
    List $10.00$1.90
    Cache read
    List $0.10$0.019
    Cache write
    List $2.50$0.47
  • claude-haiku-5-5
    Input
    List $0.10$0.019
    Output
    List $0.50$0.095
    Cache read
    List $0.010$0.002
    Cache write
    List $0.13$0.024

Why it costs less

Our Claude API comes from a pool of official subscription accounts and is billed at the official list price times one discount.

Tools like Claude Code resend the same context on every turn. Tokens read from the cache are billed at the official cache read rate, far below the input rate.

Sonnet 5.5 input list price, per 1M tokens

Cache miss$2.00
Cache hit$0.10

Anthropic pricing page

Every call on the record

Your usage log lists the model, token counts, list price, and your price for each call.

Model
The model used for each call.
Usage
Input, output, and cache tokens, itemized.
Cost
List price and your price, side by side.
How to check model and usage (in Chinese)

Set up in three steps

Read the docs
  1. 1

    Create an API key

    Sign up and create a key in the console.

  2. 2

    Set your base URL

    Point your tool at https://1zheai.com and paste your key.

  3. 3

    Make your first call

    Check its usage and cost in the usage log.

~/.zshrc
export ANTHROPIC_BASE_URL=https://1zheai.com
export ANTHROPIC_AUTH_TOKEN=YOUR_API_KEY

Top-ups and refunds

Minimum top-up¥100
Payment
WeChat PayAlipayStripeUSDT
Exchange rateConverted to USD balance at the real-time rate
RefundsUnused credit, to the original payment method, within 7 days of top-up
InvoicesWeChat Pay and Alipay: fapiao for an extra 6% of the payment amount. Stripe: Stripe invoice. USDT: none.

The docs cover Claude Code, Cursor, and the Anthropic Python and Node.js SDKs. Other tools need to support a custom base URL.

Start with ¥100

Create an API key in the console, then follow the docs to connect your tools.