---
title: Usage and costs
description: Understand input, output, conversation history, and where to check actual request charges.
lastUpdated: 2026-09-08
---

Infistar charges for actual usage. Prices and billing units differ by model; consult the current [Model Square](https://infistar.ai/pricing) before use.

## Input and output

A text conversation usually includes both input (your question, uploaded content, and conversation history) and output (the generated answer). Both can contribute to usage. The selected model's current rate and billing unit determine the charge.

## Why longer conversations can cost more

Clients commonly resend previous messages when you ask a follow-up. As the conversation grows, the input sent with a new request may grow too. Start a new conversation when you no longer need the earlier context.

## Images and files

Their billing depends on the selected model and actual input. Do not apply a text model's token price to an image or video model, or assume every media model uses the same billing unit. Use the model's current pricing description and the actual usage record.

## Check your usage

In **Console → Usage Logs**, inspect request time, model name, input and output usage, and actual consumption. **Wallet Management** shows the balance and top-up records.

To control costs, choose a suitable lower-cost model for simple tasks, trim unnecessary file content, start a new conversation when appropriate, and disable keys you no longer need. Do not rely on fixed prices copied from an old tutorial.

For adding credit, see [Top-ups and redemption](/en/getting-started/recharge-finance).