Billing & Tokens

INFER uses a prepaid token-based billing system. Buyers add funds to their account and pay per token of inference used. Operators earn a share of each request processed.

How Billing Works

For Buyers

1. Add funds — Deposit USDC or pay with credit card to add balance to your account
2. Use the API — Each inference request deducts tokens from your balance based on usage
3. Monitor usage — View detailed usage logs in the dashboard with per-request breakdowns

For Operators

1. Set pricing — Configure your price per 1K tokens in the operator dashboard
2. Serve requests — Process inference requests routed to your node
3. Earn tokens — Receive 80% of the request cost as earnings
4. Withdraw — Transfer earnings to your wallet or bank account

Pricing Model

Pricing is based on tokens processed — both input (prompt) tokens and output (completion) tokens.

Default Platform Pricing

Operators can set custom pricing that overrides the platform defaults.

Revenue Split

Token Counting

Tokens are counted using the model’s native tokenizer. As a rough approximation:

• 1 token ≈ 4 characters in English
• 1 token ≈ ¾ of a word
• 100 tokens ≈ 75 words

Both input and output tokens are metered. The total cost for a request is:

cost = (input_tokens × input_price) + (output_tokens × output_price)

Balance Management

Adding Funds

Funds can be added through:

• Crypto — Send USDC to your INFER deposit address (Ethereum, Base, or Solana)
• Credit Card — Pay with Visa/Mastercard (processed via Stripe)

Minimum deposit: $5.00. No maximum.

Low Balance Warnings

The platform sends notifications when your balance drops below:

• $5.00 — Email warning
• $1.00 — Email + desktop app notification
• $0.00 — API requests return 402 Payment Required

Earnings Withdrawal

Operators can withdraw earnings when:

• Minimum balance of $10.00 in earnings
• Account is in good standing (no violations)
• KYC verified (for withdrawals over $500)

Usage Logs

Every API request is logged with:

• Timestamp
• Model used
• Input and output token counts
• Cost breakdown
• Node that served the request
• Response latency
• Status (success/error)

View usage logs at Dashboard → Usage Logs or via the API at GET /api/usage.

Billing FAQ

Q: What happens if my balance runs out mid-request?

The request will complete normally. Your balance may go slightly negative (up to the cost of one response). Add funds to continue making requests.

Q: Are there volume discounts?

Not currently. All users pay the same per-token rate. Enterprise pricing with volume commitments is planned for Q3 2026.

Q: How are streaming requests billed?

Streaming requests are billed identically to non-streaming requests — based on total input + output tokens. The streaming format doesn’t affect pricing.