Compare per-request and monthly API costs across OpenAI, Anthropic and Gemini, with the price check date.
Estimate the VRAM an LLM needs and check whether it fits on your GPU.
Open →Paste a document to see RAG chunking and embedding cost estimates live. Your text is never uploaded.
Open →Convert between world currencies, several at once. Rates update daily and are for reference only.
Open →No comments yet. Be the first!
The LLM API Cost Calculator turns tokens per request and requests per day into per-request, monthly and yearly cost, and compares several OpenAI, Anthropic and Google Gemini models side by side. Formula: request cost = uncached input × input price + cache-read × cache price + cache-write × write price + output × output price (all prices per million tokens). It supports prompt caching, Batch/Flex/Fast, long-context pricing and custom models, with price sources and the verification date shown under the tool.
Request cost = uncached input tokens × input price + cache-read tokens × cache price + cache-write tokens × write price + output tokens × output price, with every price quoted per million tokens. Multiplying by requests per day and days per month gives the monthly and yearly cost.
No. Prices are static data kept in the site's code and updated by hand against each vendor's official pages. The sources and a "verified on" date are listed under the tool, and a "may be out of date" warning appears after 30 days. You can also add a custom model with your own quote.
"Share served from cache" is the part of each request's input read from the cache. Anthropic also has a "share written to cache", which usually costs more than ordinary input. OpenAI and Gemini have no separate write price, so written tokens are billed as ordinary input. The two shares cannot add up to more than 100%.
Batch and Flex trade a lower price for slower or queued processing, while Fast pays more for quicker output. Each vendor supports different modes and multipliers; a model that does not support your chosen mode is listed as skipped. Where OpenAI's page only says "about 50%" or "about 2×", the table marks the result as approximate.
Some models switch the whole request to a higher rate once a single prompt passes a certain length. The threshold differs by model and is applied per request, so enter the input tokens of ONE request, not the monthly total. Models priced at the long-context rate are labelled in the result table.
The most accurate source is the usage field in each API response or the vendor's own token-counting endpoint. This tool's text estimate gives only a range (about 4 characters ≈ 1 token in English, about 1 character ≈ 1 token in CJK), and tokenizers can differ noticeably; newer Claude models produce about 30% more tokens for the same text. Remember that output tokens include reasoning (thinking) tokens.
Usual reasons: add-on charges such as tool calls and web search, reasoning tokens, tax, separate cloud-marketplace pricing (Bedrock, Vertex and similar), regional uplifts, negotiated or committed-use discounts, and real token counts that differ from your estimate. This tool only calculates token cost at public standard prices.
No. Everything is calculated in your browser. Pasted text is only used to estimate tokens locally and is never sent to a server, and no paid API or model provider is called.
Paste a sample prompt or reply for a rough range (not an exact count from any vendor tokenizer).
OpenAI
Anthropic
Google Gemini
| Model | Per request | Per 1,000 | Per month | Per year | Output share |
|---|---|---|---|---|---|
Claude Haiku 5.5Lowest new tokenizer: about 30% more tokens for the same text | $0.000450 | $0.450 | $13.50 | $162.00 | 56% |
Gemini 2.5 Flash audio input is priced differently; text / image / video prices are used here | $0.00185 | $1.85 | $55.50 | $666.00 | 68% |
GPT-5.4 mini | $0.00375 | $3.75 | $112.50 | $1,350.00 | 60% |
Gemini 2.5 Pro | $0.00750 | $7.50 | $225.00 | $2,700.00 | 67% |
Claude Sonnet 5.5 new tokenizer: about 30% more tokens for the same text | $0.00900 | $9.00 | $270.00 | $3,240.00 | 56% |
GPT-5.4 | $0.0125 | $12.50 | $375.00 | $4,500.00 | 60% |
Price sources and verification
Data verified on 2026-10-20
This tool is not affiliated with OpenAI, Anthropic or Google. Prices are each vendor's public standard API list prices in USD, excluding tax, negotiated pricing, cloud-marketplace pricing (Bedrock, Vertex, etc.) and add-ons such as tool calls or web search.
