LLM cost calculator: API pricing for 340+ models

Every major LLM API bills by the token: small chunks of text that average about four characters in English. Input tokens (everything you send: system prompt, conversation history, retrieved documents) and output tokens (the model’s reply) have separate prices, quoted in US dollars per million tokens. This calculator applies those published prices to your workload. To get real numbers, count the tokens in a real prompt.

For one request, the cost is:

(uncached input × input price + cached input × cached price + output × output price) ÷ 1,000,000

The daily cost multiplies that by your requests per day. The monthly figure uses the average month length of 365 ÷ 12 = 30.4 days, and the yearly figure uses 365 days. For example, a request with 2,000 input and 500 output tokens on GPT-6.1 Sol costs (2,000 × $2 + 500 × $10) ÷ 1,000,000 = $0.009.

How much does the OpenAI API cost?

OpenAI’s 57 paid models in our data cost from $0.018 to $150 per million input tokens (as of 16 October 2026), and output costs more. The cheapest by blended price is gpt-oss-20b. See every model and price on the OpenAI pricing page, or pick one in the calculator.

How much does the Claude API cost?

Anthropic’s 15 paid models in our data cost from $0.10 to $15 per million input tokens (as of 16 October 2026), and output costs more. The cheapest by blended price is Claude Haiku 5.5. See every model and price on the Anthropic pricing page, or pick one in the calculator.

How much does the Gemini API cost?

Google’s 21 paid models in our data cost from $0.05 to $2 per million input tokens (as of 16 October 2026), and output costs more. The cheapest by blended price is Gemma 3 4B. See every model and price on the Google pricing page, or pick one in the calculator.

How much is 1,000 or 1 million tokens?

Prices are quoted per million tokens, so 1 million tokens costs the listed price and 1,000 tokens costs a thousandth of it. For the popular models below, 1,000 input tokens cost from $0.0001 (GPT-6 Luna) to $0.004 (Claude Opus 5.5). The prices table lists the per-1,000 figure for each.

Why are output tokens the expensive part?

Output is priced at 3× to 5× the input price for the popular models below, because each output token is generated one after another. Two things inflate output more than people expect. First, reasoning models bill their hidden “thinking” as output tokens, so a short visible answer can still cost thousands of tokens. Second, a verbose system prompt (“explain your reasoning step by step”) makes every reply longer. Asking for concise answers is often the cheapest optimisation available.

How much does prompt caching save?

If many requests begin with the same text, providers can keep that prefix in a cache and charge 2% to 27% of the normal input price to read it again. The calculator assumes the cache is already warm. Some providers also charge a one-time premium to write to the cache, which matters if your prefix changes often. Caching only applies to an identical prefix, so put the fixed parts of your prompt first and the parts that change last. To see how often your traffic will actually hit the cache, and whether the write premium pays off, see what prompt caching saves.

When do batch and long-context prices apply?

Batch APIs run requests in the background, usually within 24 hours, at roughly half price. When you turn on batch mode, the calculator uses each provider’s published batch price, and it tells you if a model has none. To compare batch API pricing with real-time cost for a mixed workload, with each provider’s turnaround and limits, use the batch calculator. Some models also switch to a higher rate once a prompt passes a size threshold, such as 200,000 tokens. When that happens the higher rate covers the whole request, and the calculator shows a note.

Which LLM API is cheapest for this workload?

The alternatives table ranks every model whose context window and output limit can handle your request, from cheapest up. It doesn’t judge quality: the cheapest model may not be good enough for your task. Use it to shortlist candidates, then test them on real prompts before switching. For an overview, see the cheapest LLM APIs; if you chat more than you build, also compare with ChatGPT Plus or Claude Pro.

Prices come from the OpenRouter models API and the open-source LiteLLM price list and are refreshed daily (last updated 2026-10-16). For open-weight models such as Llama, the price is a typical hosted price; running them on your own hardware costs something else entirely. The methodology page explains how the data is collected and checked, and the AI pricing changes log shows which prices moved recently.