Enter token counts and your model's price per 1M tokens to estimate a call's cost. Works for any provider — plug in whatever rate your model charges. Don't know your token count? Use the AI Token Counter first.
Check your provider's pricing page (OpenAI, Anthropic, Google, etc.) — prices vary by model and change over time, so this tool deliberately doesn't hardcode them. Most providers list price per 1M tokens directly, or per 1K — multiply a per-1K price by 1000 to get the per-1M figure this tool expects.
Nearly all LLM APIs charge a higher rate for output (generated) tokens than input (prompt) tokens, so splitting them gives a more accurate estimate.
It's a straightforward multiplication of what you enter — accurate as long as your token counts and rates are accurate. It doesn't account for caching discounts, batch pricing, or provider-specific rounding.