Free LLM pricing calculator
Know the real cost.
Before you generate.
Paste a prompt and expected output. See exactly what Claude, GPT, Gemini, Grok, DeepSeek and 12 other providers would charge, side by side, by speed tier and context length.
No signup · 17 providers · 60+ models
Example length
Quantity
pieces
| Model | Company | Input $/MTok | Output $/MTok | Input cost | Output cost | Total cost |
|---|
Token counts are estimated at ~4 chars/token (English text). Actual counts depend on each model's tokenizer and may differ ±15–20%. * Price sourced from third-party aggregator — verify before relying on it. Sources: Claude · Z.ai · OpenAI · xAI · Google · DeepSeek · Kimi · Mistral
Frequently asked questions
What is llm-test.com?
llm-test.com is a free, independent LLM pricing calculator. You paste a prompt and the length of the response you expect, and it shows exactly what Claude, GPT, Gemini, Grok, DeepSeek and 12 other providers would charge for that text — side by side, by speed tier and context length. It covers 17 providers and 60+ models, needs no signup, and is not affiliated with any model provider.
How is LLM API pricing calculated?
LLM APIs bill per token, not per request. You pay separately for input tokens (your prompt plus any context) and output tokens (the model's response), each at a per-million-token rate that differs by model. The cost of a call is (input tokens ÷ 1,000,000 × input rate) + (output tokens ÷ 1,000,000 × output rate). llm-test.com runs this calculation across every model at once.
What is a token, and how many tokens are in a word?
A token is the unit of text an LLM reads and generates — a chunk of roughly 4 characters, or about three-quarters of a word in English. So 1,000 tokens is around 750 words, and 100 words is about 130 tokens. Because providers price and bill in tokens, token count rather than word count determines what you pay; exact counts vary ±15–20% by each model's tokenizer.
Why do output tokens cost more than input tokens?
On almost every price list, output tokens cost 3–8× more than input tokens. Input is processed in parallel in a single forward pass, while output is generated one token at a time, each requiring its own full pass through the model — so output is far more compute-intensive. Read the full explanation →
Which LLM is the cheapest?
There is no single cheapest model — it depends on how long your prompt and response are and which speed tier you need. For short, high-volume tasks the lowest per-token models (budget tiers from DeepSeek, Google Gemini Flash or open models) usually win; for long responses the output rate dominates. llm-test.com highlights the cheapest option for your exact input in green so you can compare before you commit.
How much does it cost to run an LLM like GPT, Claude or Gemini?
It depends on volume and model. A single short answer of a few hundred tokens typically costs a fraction of a cent, but cost scales linearly with tokens and with the number of calls — 1,000 calls cost 1,000× one call — and a flagship model can cost 10–50× more per token than a small one. Enter your real prompt and expected output into the calculator to see the actual figure for each model.