About this tool
LLM features live or die on unit economics, and the pricing pages make the math awkward: prices per million tokens, split between input and output, while your product thinks in requests per day. This calculator bridges that. Pick a model from the built-in October 2026 price list covering the mainstream OpenAI, Anthropic and Google lineups — from GPT-5.5 and Claude Opus 5.5 to Gemini 3 Pro and the budget mini/nano and Flash tiers, enter your average input and output tokens per request and your daily request volume, and get an instant monthly cost breakdown showing input cost, output cost and total. Compare two scenarios side by side by tweaking the numbers — see what caching-friendly short prompts save, or what a model downgrade does to your margin at 10,000 requests a day. A per-request figure helps you sanity-check pricing against what users pay. All arithmetic happens in your browser; your usage projections stay private. Prices are compiled from public rate cards and labeled clearly, but vendors change pricing often — always confirm against the official pricing page before committing to a budget.
How to use LLM API Cost Calculator
- Select a model from the price-list dropdown.
- Enter average input and output tokens per request.
- Enter your expected requests per day.
- Read the estimated monthly cost, split by input and output tokens.
Frequently asked questions
Are the prices current?
They reflect public rate cards as of October 2026 (GPT-5.5, Claude Opus/Sonnet/Haiku 5.5, Gemini 3 Pro and 2.5 series) and are shown in the tool. Vendors reprice often, so verify on the official pricing page before budgeting.
How do I estimate tokens per request?
A rough guide: 750 English words is about 1,000 tokens. Use our Context Window Calculator to convert your typical prompt and response lengths.
Does it include discounts like caching or batch?
No, it uses list prices. Cached-input and batch-API discounts can cut real bills substantially for suitable workloads.
Is my usage data sent anywhere?
No. The calculation is pure local arithmetic; nothing is transmitted.