AI cost calculator
Compare monthly API costs across 10 frontier models. Sortable, filterable, instantly updated. No signup, no upload.
| Model | Total / mo ↓ |
|---|
Costs are without prompt-caching or batch-API discounts. See cache calculator and batch calculator for those.
How this calculation works
Plug in three numbers: input tokens per call, output tokens per call, and calls per month. We compute monthly cost for each model using their published per-million-token rates, then sort by total.
Don't know your token counts? Use the token counter to get exact numbers from a sample prompt.
Why your real bill is usually lower
These figures show list prices. Real production costs are usually lower because of prompt caching (Anthropic ~90% off cached input), batch APIs (OpenAI 50% off), or volume commitments. See:
- Prompt-caching ROI calculator: model what your bill drops to once you turn caching on
- Batch API breakeven calculator: when async batch is worth the latency tradeoff
- Agent-loop cost simulator: for multi-turn tool-using agents where context grows turn-by-turn