Credits System

CodingFleet uses a flexible credit-based system that adapts to your plan. Credits determine how much usage you get over time.

1. Credit Types by Plan

a) Free & Pro Plans

• You receive limited but permanent credits.
• Free users get 10 credits per month.
• Example:
  • Pro ($10/month) → 200 credits per month
• Monthly credits:
  • Are added to your account
  • Do not expire
  • Are consumed as you use models
• Once credits run out, you can no longer use models until you add more credits.

Purchasing Extra Credits

• Purchased credits are always permanent, regardless of your plan.
• If you’re on the Free plan, purchasing credits also grants a 30-day Pro pass.
• After 30 days, the Pro access ends, but your credits remain.

b) Unlimited, Elite & Ultimate Plans

These plans use a weekly quota system, not permanent monthly credits.

• Weekly credits reset every Monday midnight UTC time.
• When you exhaust your weekly quota:
  • You are automatically downgraded to unlimited fast models (e.g. GPT-5 mini, Grok-4.1 Fast).
  • These models remain fully unlimited for you.

Using Premium Models After Quota

• To continue using premium models (Claude Sonnet/Opus, Gemini Pro, GPT, etc.):
  • You can purchase additional permanent credits
  • Purchased credits are consumed after your weekly quota
  • Unused purchased credits never expire
  • Purchasing credits is always permanent, no matter your plan.

2. How Credit Costs are Calculated

Each model has a base cost, shown when selecting it (e.g. -5, -7). Overall response cost depends on total tokens consumed (1 token is roughly 3-4 characters)

Base Cost Rule

• If a request uses less than 50k total tokens (input + output):
  • Cost = base cost

Large Requests (Over 50k Tokens)

• Once usage exceeds 50k tokens:
  • Cost scales per 25k tokens
• Formula: (total_tokens // 25,000) × base_cost
• Example:
  • Model base cost: 7
  • Input tokens: 60k
  • Output tokens: 30k
  • Total: 90k tokens
  • 90k // 25k = 3
  • 3 × 7 = 21 credits
  • ➡️ Final cost: 21 credits
• Sometimes when there are cache hits (usually occurs on quick successive prompts in the chat tool), the cost is reduced up to 90% of those tokens. For example, if total tokens consumed are 90k, but 50k are cache hits, then total cost will be calculated as if it's only 90k - 50k + 50k * 0.1 = 45k tokens.

3. Key Takeaways

• Free & Pro → permanent credits, limited usage
• Unlimited+ → weekly quota + unlimited fallback models
• Purchased credits never expire
• Costs scale automatically for large prompts and responses

   Last Updated: dec 2025