Pay only for what you use
No subscription. No markup on model rates. You pay the same per-token prices your providers charge, metered by request.
No markup · No subscription · Prepaid balance, never expires
Model pricing, passed through
Rates for popular models across every provider, per 1M tokens. Infere adds no markup - the live registry is the source of truth.
| Model | Provider | Input / 1M | Output / 1M | Cached input / 1M |
|---|---|---|---|---|
| gpt-5.6-luna | OpenAI | $0.20 | $1.20 | $0.02 |
| gpt-5.4 | OpenAI | $2.50 | $15.00 | $0.25 |
| o3 | OpenAI | $2.00 | $8.00 | $0.50 |
| Claude Haiku 4.5 | Anthropic | $1.00 | $5.00 | $0.10 |
| Claude Sonnet 5 | Anthropic | $2.00 | $10.00 | $0.20 |
| Claude Opus 4.6 | Anthropic | $5.00 | $25.00 | $0.50 |
| DeepSeek V4 Pro | Fireworks AI | $1.20 | $1.20 | $0.60 |
| GLM 5.3 | Fireworks AI | $1.40 | $4.40 | $0.26 |
| Llama 3.3 70B Instruct Turbo | Together AI | $1.04 | $1.04 | - |
| DeepSeek V4 Pro 0813 | Together AI | $1.32 | $3.96 | $0.13 |
| gemma-4-31B-it | Infercom | $0.20 | $0.35 | - |
| gpt-oss-120b | Infercom | $0.22 | $0.59 | - |
A representative sample; rates change when providers change their pricing, and some models vary by context length or service tier. Verify with the provider, or open the live registry for every model from OpenAI, Anthropic, Fireworks AI, Together AI, and Infercom.
How pricing works
Three steps. Start free, add balance only when you are ready.
1. Start free
Create your account and API token at no cost. No card required to begin building.
2. Add balance
Prepay any amount through Stripe hosted checkout, whenever you are ready. Your balance stays until you use it.
3. Pay per request
Each request is charged at the model's provider rate, based on input, output, and cached-input tokens, deducted in real time.
Controls to stay on budget
Spend is capped and visible from day one.
Token budgets
Cap spend per token with daily, weekly, or monthly limits and a warning threshold. A hard limit blocks requests; a soft limit flags them.
Auto-recharge
Top up automatically when your balance drops below a threshold, with daily and monthly caps. Off by default.
Usage alerts
Get notified as balances run low or budgets approach their limits.
Cost analytics
Break down spend by model, provider, and user, with trends and exports.
Frequently asked questions
How does pricing work?
Infere has no subscription and adds no markup to model rates. You prepay a balance, and each request is charged at the provider's per-token rate for the model used, deducted in real time.
Do I have to pay to start?
No. You can create an account and generate an API token for free. Add a balance only when you are ready to send requests.
How are model costs calculated?
Each request applies the model's input, output, and discounted cached-input rates to the tokens used. Rates are shown per 1M tokens, and the in-app Model Registry (/models) lists every current rate. Those values are the source of truth.
What happens when my balance runs out?
The gateway rejects the request with an insufficient-credit error (402). Enable auto-recharge or top up manually in Settings → Billing → Credits.
Is there a subscription or minimum spend?
No subscription and no minimum. You only pay for the tokens you use, and any unspent balance stays on your account.
Start free, pay only for what you use
No subscription, no markup on model rates. Add balance only when you are ready.