Cardwave

Chat turn budget calculator

Chat costs grow as a conversation lengthens, because every turn resends the history so far. Enter a model's per-token prices and your typical turn sizes to see how many turns a budget buys, and where the context window caps the cost.

A turn here means one user message plus the one AI response to it.

Input cost / M tokens ($)

Output cost / M tokens ($)

Max budget (USD $)

System prompt (tokens)

User input (tokens)

Model response (tokens)

Max context size (tokens)

Metrics

Max turns —

Final cost —

Window capping point—

Cost of first turn—

Cost of capped turn—

Sliding window: the system prompt is a fixed baseline cost on every turn. Once history fills the remaining context window, the input size stops growing and turn costs flatten out.