LLM cost calculator
Enter your projected monthly token volume to see the cost across all models.
Values are in millions of tokens. "10" means 10M tokens/month.
| Rank | Model | Provider | Input | Output | Monthly total | Free |
|---|
※ Cache pricing and batch discounts not included. Actual cost varies by usage pattern.
Estimating your token volume
- Rule of thumb: 1M tokens is roughly 750,000 English words. Japanese and Chinese text usually needs more tokens per character, and the exact ratio depends on each model’s tokenizer.
- Output tokens typically cost 3–8× as much as input tokens. Chat and agent workloads with long answers are dominated by output cost; RAG, classification and extraction are dominated by input cost.
- Not included here: prompt caching (cached input is often 75–90% cheaper) and batch APIs (usually half price for asynchronous jobs). Model pages list cache and batch prices where available.