Skip to main content

​
Symptom Description

When calling the API, you receive the following error messages: 
{
  "error": {
    "message": "Insufficient balance",
    "type": "insufficient_balance",
    "code": "insufficient_balance"
  }
}
Or similar prompts:
  • “Account balance is insufficient”
  • “Please recharge”
  • “Balance insufficient, unable to complete request”

​
Cause Analysis

​
1. Balance Depleted

Account balance has been completely used up and needs recharge.

​
2. Insufficient Balance

Account has remaining balance, but insufficient for current request:
  • Request token count is too large
  • Selected model is expensive
  • Single request exceeds budget

​
3. Frozen Balance

Part of balance is frozen and unavailable:
  • Pending transactions
  • Risk control freezing
  • System reservation

​
4. Billing Method Mismatch

Using wrong billing method token:
  • Pay-per-use token calling usage-based billing models
  • Usage-based billing token calling pay-per-use models

​
Solutions

​
Immediate Solutions

1

Check Account Balance

  1. Log in to Laozhang API Console
  2. Check dashboard balance display
  3. View detailed transaction history
2

Recharge Account

  1. Click “Recharge” button in console
  2. Select recharge amount
  3. Complete payment process
  4. Wait for account balance update (usually instant)
3

Verify Recharge Success

  1. Refresh console page
  2. Check if balance has updated
  3. Retry API call

​
Emergency Alternatives

If urgent need to continue using:
  1. Switch to Lighter Model 
    # Original: Using expensive model
model = "gpt-4-turbo"

# Alternative: Switch to economical model
model = "gpt-3.5-turbo"
  2. Reduce Request Parameters 
    # Reduce max_tokens
response = client.chat.completions.create(
    model="gpt-4-turbo",
    max_tokens=500,  # Reduce from 2000 to 500
    messages=[...]
)
  3. Use Backup Account
    • Switch to alternative API key
    • Switch to different service provider

​
Preventive Measures

​
1. Set Low Balance Alert

In console set alert threshold:
  • Recommended setting: 20% of usual daily usage
  • Alert methods: Email, SMS, webhook
  • Check frequency: Daily automatic check

​
2. Enable Auto-recharge

Configure auto-recharge rules: 
Trigger condition: Balance < $10
Recharge amount: $50
Payment method: Credit card auto-deduct

​
3. Budget Management

Set usage budget:
  • Daily limit: Prevent unusual high usage in single day
  • Monthly limit: Control overall cost
  • Model-specific limit: Limit expensive model usage

​
4. Monitor Usage

Regularly check usage: 
Weekly tasks:
- Check balance trend
- Analyze usage distribution
- Identify unusual usage
- Optimize cost structure

​
5. Choose Appropriate Billing Method

Choose based on usage pattern:
Usage PatternRecommended MethodReason
Stable, high-frequencyUsage-based billingLower unit price
Occasional usePay-per-useNo waste
Image/Video generationPay-per-useClear pricing
Chat applicationsUsage-based billingMore economical

​
Balance Management Best Practices

​
Budget Allocation

Reasonably allocate budgets: 
Total monthly budget: $100

Allocation plan:
- Production environment: $60 (60%)
- Development testing: $20 (20%)
- Emergency reserve: $20 (20%)

​
Cost Optimization

Reduce unnecessary expenses:
  1. Model Selection Optimization 
    # Simple tasks use economical models
if task_complexity == "simple":
    model = "gpt-3.5-turbo"
else:
    model = "gpt-4-turbo"
  2. Enable Caching 
    # Cache common requests
from functools import lru_cache

@lru_cache(maxsize=100)
def get_completion(prompt):
    return client.chat.completions.create(...)
  3. Batch Processing 
    # Batch process requests to reduce overhead
results = []
for batch in chunks(requests, batch_size=10):
    results.extend(process_batch(batch))

​
Usage Tracking

Record and analyze usage: 
import logging

# Log each API call
logging.info(f"API call: model={model}, tokens={tokens}, cost=${cost}")

# Regularly generate usage reports
def generate_usage_report():
    """Generate weekly usage report"""
    total_cost = sum(costs)
    total_requests = len(costs)
    avg_cost = total_cost / total_requests
    
    print(f"Total cost this week: ${total_cost}")
    print(f"Total requests: {total_requests}")
    print(f"Average cost per request: ${avg_cost}")

​
Common Questions

Recharge arrival time:
  • Online payment (credit card, PayPal): Usually instant
  • Bank transfer: 1-3 business days
  • Cryptocurrency: Wait for blockchain confirmation, about 10-30 minutes
If delayed:
  1. Check payment status
  2. Confirm correct account information
  3. Contact customer support
Refund policy:
  • Balance can be refunded without violations
  • Refund processing time: 3-7 business days
  • May deduct processing fees (typically 3%-5%)
Refund process:
  1. Submit refund request in console
  2. Provide payment proof
  3. Wait for review
  4. Receive refund
Estimation methods:
  1. Token Count Estimation
    • English: ~1 word = 1.3 tokens
    • Chinese: ~1 character = 2 tokens
  2. Use Official Pricing Calculator Visit Pricing Page for calculation
  3. Reference Historical Usage View usage in console
Example calculation:
Request: 1000 token prompt
Response: 2000 token completion
Model: gpt-4-turbo ($10/M tokens)

Cost = (1000 + 2000) / 1,000,000 * $10 = $0.03
Prevention measures:
  1. Set Rate Limits 
    from ratelimit import limits, sleep_and_retry

@sleep_and_retry
@limits(calls=10, period=60)  # Limit 10 requests per minute
def call_api():
    return client.chat.completions.create(...)
  2. Implement Request Validation 
    def validate_request(prompt):
    token_count = estimate_tokens(prompt)
    if token_count > 10000:
        raise ValueError("Request too large")
  3. Enable Budget Alerts Set daily/weekly/monthly budget alerts
  4. Code Review Regularly review API call code to prevent loops or repeated calls
Free Tier Policy:Laozhang API currently does not offer free tier, but provides:
  • New user discount: First recharge gets 10% bonus
  • Volume discount: Large recharge gets higher discounts
  • Promotional activities: Occasional promotional offers
Cost Reduction Recommendations:
  1. Use economical models (GPT-3.5 Turbo)
  2. Optimize prompt length
  3. Enable result caching
  4. Batch process requests

​
Emergency Contact

If you cannot resolve the issue through above methods, please contact us through:
  • Online Support: Click chat icon in console
  • Email Support: [email protected]
  • Work Hours: Monday to Friday 9:00-18:00 (UTC+8)
  • Emergency Contact: For production environment issues, specify “Emergency” in email subject
⌘I