Skip to main content

​
Deposit funds

LLMTune uses a prepaid balance (your “wallet”). You add funds up front, then inference, fine-tuning, embeddings, and the Agent are deducted as you go. There are no end-of-cycle invoices — you only pay for what you use, in real time.
This prepaid wallet is separate from Coding Plans, which are monthly subscriptions that include a fixed quota of AI requests for the CLI and IDEs. Wallet balance is only charged for usage outside your plan’s quota — non-plan models, fine-tuning, and embeddings.

​
Deposit funds

  1. Sign in to the platform at https://llmtune.io/login.
  2. Open Billing from the left sidebar (or go to https://llmtune.io/billing).
  3. Your current balance is shown at the top of the page.
  4. Click Add funds (or Top up).
  5. Choose your payment method.
  6. Enter the amount you want to deposit.
  7. Confirm the payment. Your balance updates immediately on success.

​
How balance is used

WorkloadHow it charges
Inference (/v1/chat/completions)Per request, based on input + output tokens.
Embeddings (/v1/embeddings)Per request, based on input tokens.
Fine-tuning (/api/training)Per GPU hour, billed while the job runs (not while queued).
AgentPer request / per step, like inference.
Each successful call is deducted from your balance. Failed requests are not charged — if a request errors (4xx/5xx), no tokens are billed.

​
What happens when balance runs out

When your balance can’t cover a request, the API returns 402 Payment Required: 
{
  "error": "Insufficient balance",
  "message": "Insufficient balance. Required: $0.00, Available: $0.00",
  "code": "INTERNAL_ERROR",
  "model": "qwen/qwen-2.5-7b-instruct-confidential"
}
To recover:
  1. Go to Billing → Add funds.
  2. Deposit enough to cover the request (the Required field shows the minimum).
  3. Retry your request. No data is lost — the request simply didn’t run.
Even models marked Confidential (the free / own-your-data tier) go through the paid inference path. If your balance is $0.00, confidential models will also return 402. Deposit any amount to unlock them.

​
Best practices

  • Deposit a buffer — Add more than your immediate need to avoid mid-job 402s during batch runs.
  • Estimate before training — For fine-tuning, use the cost estimate in Fine-tune Studio before you launch, and make sure your balance covers the full GPU-hours estimate.
  • Handle 402 in code — Treat 402 as recoverable: pause, prompt the user to add funds, then retry. Don’t auto-retry in a tight loop.
  • Monitor usage — Check Billing & usage to track spend and spot unexpected drains early.

​
Next