Pricing & Billing
AICredits pricing explained — USD-to-INR conversion, special token pricing for cache writes, reasoning tokens, and FIFO credit deduction.
Use this page with an AI assistant
Opens a new chat with this docs URL and the correct AICredits base URLs.
Pay-per-token billing in INR. Costs are calculated per request based on the model used and tokens consumed.
How Pricing Works
Each model has separate rates for input (prompt) tokens and output (completion) tokens. After each request, AICredits calculates the cost and deducts it from your wallet.
Token costs are based on each model's published per-token rates, converted to INR at the current USD/INR rate. The final INR price you see in the cost calculator and on your usage page is exactly what gets deducted from your wallet — no hidden charges.
The current exchange rate is shown on your Dashboard. Prices are locked at the rate active at the time of each request.
INR Conversion
All model costs are denominated in USD by providers. AICredits converts them to INR using a live exchange rate that is refreshed periodically and visible on your Dashboard. A small buffer is included to account for rate fluctuations between request time and settlement.
Special Token Pricing
Some providers have non-standard token pricing for caching and reasoning:
| Provider | Token Type | Multiplier |
|---|---|---|
| Anthropic (Claude) | Cache writes | 1.25× base input price |
| Cache reads | 0.1× base input price | |
| Long context (>200K tokens) | 2× input price | |
| OpenAI | Cached input | 0.5× input price (50% discount) |
| Reasoning tokens | Billed as output tokens | |
| Google (Gemini) | Cache writes | 0.25× base price |
| Thinking tokens | Billed separately |
Pricing Tiers
Some models support multiple pricing tiers with different cost and latency trade-offs:
| Tier | Description |
|---|---|
| Standard | Default pricing with standard latency |
| Batch | Lower cost for non-time-sensitive tasks (higher latency) |
| Flex | Dynamic pricing based on demand |
| Priority | Premium pricing for lowest latency and highest availability |
FIFO Credit System
Credits are consumed in first-in, first-out order. When you add credits, they are stored as ledger entries with a 1-year expiration. Oldest credits are used first. This ensures you always use your earliest credits before they expire.
Checking Your Balance
Check your balance via the Dashboard or programmatically:
curl https://api.aicredits.in/v1/credits \
-H "Authorization: Bearer sk-your-key-here"
# Response:
# {
# "data": {
# "total_credits": 10.50, ← USD equivalent
# "total_usage": 0,
# "credits_inr": 955.50 ← Actual INR balance
# }
# }Low Balance
A low balance alert triggers when your wallet drops below ₹10. Ensure you maintain sufficient credits to avoid request failures.