Skip to main content

​
APIYI Pricing Principles

​
1. Model Price Alignment

  • Overseas Models: Prices equal to official websites (including OpenAI, Claude, Gemini, Grok, etc.)
  • Domestic Models: Prices lower than official websites (including DeepSeek, Qwen, Doubao, Kimi, etc.)
  • Rare Models: Some newly released overseas models (e.g., o3-pro) or models with high access thresholds may have slight markups. However, with top-up bonuses and discounts, prices are close to official rates.

Transparent Pricing, 20% Off

Most model prices are consistent with official rates. Consumption is completely transparent, just for peace of mind~

Complete Models, Fast Updates

Whenever manufacturers release new models, APIYI updates promptly. Speed is an advantage!

​
Top-up Advantages

​
2. Exchange Rate Benefits

Preferential Exchange Rate

1:7 preferential exchange rateLower than real-time USD exchange rate (about 7.2)

Low Entry Threshold

$5 minimumJust 35 RMB to start using

​
3. Top-up Bonuses

The more you top up, the more you saveTop up $100, get 15% extra bonus, overall discount up to 20% off official prices

​
Discount Calculation Example

Using a $100 top-up as an example (overseas top-up thresholds are high, channel fees and agent markups are often significant): 
Actual Payment: 700 RMB
Credit Received: 100 + 15 = 115 USD
Effective Rate: 700 ÷ 115 = 6.08
Discount Ratio: 6.08 ÷ 7.5 = 0.81 (about 19% off)

​
More Advantages Over Official Websites

More importantly, compared to official single accounts, APIYI provides more comprehensive services:

Unlimited Usage

  • No rate limits
  • No account ban risk
  • Pay-as-you-go billing

Complete Model Selection

  • 200+ popular models
  • One-click switching
  • Continuous updates

Easy Integration

  • Unified API interface
  • OpenAI format compatible
  • Zero migration cost

Enterprise-Grade Service

  • Professional technical support
  • Stable and reliable
  • Data security guarantee

​
Billing Basics

​
What is ‘Pay-as-you-go’?

Pay-as-you-go means you only pay for the services you actually use, without monthly or annual prepayment.

Flexible Usage

  • Pay for what you use
  • No minimum consumption
  • Start or stop anytime

Cost Control

  • Real-time consumption view
  • Set quota alerts
  • Precise to each call

​
Billing Mode Priority

APIYI supports two billing modes. When a model supports both:
Per-call billing takes priority over per-token billingIf a model supports both per-call and per-token billing, the system defaults to per-call billing.

​
Token (API Key) Settings Impact

Your billing method is also affected by token configuration:

Token-Only Billing

If token is set to “token-only billing”, even if the model supports per-call billing, it will use per-token billing

Default Setting

Tokens support all billing modes by default, system auto-selects (per-call priority)

​
‘Per-Call Billing’ Scenarios

The following types of models typically use per-call billing:
  • Image Generation
  • Video Generation
  • Special Models
Applicable Models:
  • sora_image series (reverse-engineered models)
  • flux-kontext-pro (official model)
  • DALL-E series
  • Midjourney-related models
Billing Unit: Per image
View complete model price list: APIYI Pricing

​
What are Tokens?

Tokens are the basic units for AI models to process text. Understanding tokens helps you estimate and control costs.
Token Calculation Reference
  • Chinese: 1 character ≈ 1-2 tokens
  • English: 1 word ≈ 1-2 tokens
  • 1000 tokens ≈ 750 English words ≈ 500 Chinese characters

​
Token Calculation Example

Input text: "Hello, please help me write a Python function"
Token count: About 12-15 tokens

Output text: "def hello_world():\n    print('Hello, World!')"
Token count: About 15-20 tokens

Total consumption: Input + Output ≈ 30-35 tokens

​
Prompts and Completions

In each API call, costs consist of two parts:
1

Prompt - Input Tokens

All content you send to the model, including:
  • System prompts
  • User questions
  • Context information
  • Conversation history (if any)
2

Completion - Output Tokens

Content generated by the model, including:
  • Text responses
  • Code generation
  • Structured data
Different models may have different input and output prices. Usually output tokens cost more than input tokens.

​
How to Choose the Right Model?

​
Model Selection Strategy

  • Cost Priority
  • Performance Priority
  • Balanced Choice
Suitable Scenarios: Batch processing, simple tasks, testing & developmentRecommended Models:
  • gemini-2.5-flash (fast and good)
  • gpt-4.1 (official replacement for gpt-4.5)
  • deepseek-v3 (domestic excellence, high cost-performance)
Estimated Cost: $0.1-1/million tokens
See the “Popular Models” page on the left for more information.

​
Cost Estimation Method

1

Small Sample Testing

Test with a small number of samples (5-10)
2

View Consumption Logs

Check detailed token consumption for each call in APIYI console
3

Calculate Average

Calculate average token count per call
4

Estimate Total Cost

Average tokens × Expected calls × Model unit price

Practical Advice

  1. Test with cheaper models first, verify feasibility
  2. Analyze actual consumption through APIYI backend logs
  3. Choose appropriate model based on task complexity
  4. Optimize prompts to reduce unnecessary token consumption

​
Real-time Price Inquiry

Model Price List

View real-time prices for all models
  • Per-token billing prices
  • Per-call billing prices
  • Discount comparisons

Cost Calculator (In Development)

Quickly estimate usage costs
  • Input estimated usage
  • Select models
  • Auto calculate fees

​
FAQ

Login to APIYI Console, view real-time prices for all models on the model list page.
Top-up is instant, usable immediately after successful payment.
Yes, we support official invoices. Please submit invoice application in console.
Enterprise bulk top-ups can enjoy more discounts. Please contact customer service.

​
Invoice Information

​
Invoice Process

After successful online top-up or corporate transfer, invoices are issued based on actual payment amount (all prices include tax). Customers need to provide detailed invoice information: company or university name, tax ID, etc.
1

Submit Invoice Information

Top navigation in website backend - Invoicing, self-submit invoice formSubmit Invoice Application →
2

Invoice Type

  • Can issue VAT regular invoice (special invoice needs separate tax point negotiation)
  • Invoice category: Information Technology Service Fee or Data Collection Fee
  • Can provide: Purchase list with official seal
3

Delivery Time

About 1 business day, sent via WeChat or email as electronic invoice
We have many university and enterprise customers and can accommodate various reasonable reimbursement requirements.

Get Started

Register account, top up to enjoy discounts, experience 200+ AI models