Work out what an AI model actually costs per month from your token usage, and compare the major models side by side.
Not sure? A short chat turn is roughly 300 input / 200 output tokens. A request that includes a long document can easily be 20,000 input tokens.
Prices and tools change. The model prices behind this tool were last checked on 20 July 2026. Providers revise their rates, rename models and retire tiers regularly, so treat every figure here as an estimate and confirm the current price with the provider before you commit to a budget. See our model pricing table and sources.
Estimates only. These tools run entirely in your browser — nothing you type is sent to us or stored anywhere. They are simplified models of how AI providers bill, and they exclude cached-input discounts, batch rates, image and audio tokens, taxes and minimum commitments. They are general information, not procurement or financial advice.
Almost every AI provider bills the same way: you pay per token, and you pay different rates for tokens going in and tokens coming out. A token is roughly three-quarters of a word in English. Input tokens are everything you send — your prompt, the system instructions, any documents or conversation history you include. Output tokens are what the model writes back.
Output is almost always the more expensive side, often by three to six times. That is the single most useful thing to know when estimating a bill: a workload that reads a lot and writes a little is cheap, and a workload that writes long responses is not.
Enter the tokens for a typical single request and how many of those requests you expect in a month. The tool multiplies through and shows the monthly total, split into input and output so you can see which side dominates. The comparison table underneath prices the exact same workload on every model in our table, which is usually more revealing than any benchmark — the cheapest and most expensive options for identical work can differ by fifty times.
Real bills include things this estimate leaves out: cached input (most providers discount repeated prompt prefixes heavily), batch processing (often half price for work that can wait), image and audio tokens, and any minimum commitment on an enterprise contract. Treat the number here as an upper bound for a straightforward setup, then look for those discounts.
Paste a representative prompt into the token estimator to get a number to plug in here. If you are weighing this against a flat monthly plan, the subscription vs API comparison finds the break-even point.
USD per 1,000,000 tokens. Last verified 20 July 2026. These figures are maintained by hand and will drift — verify with the provider before budgeting.
| Model | Input / 1M | Output / 1M |
|---|---|---|
| Anthropic | ||
| Claude Fable 5 | $10.00 | $50.00 |
| Claude Opus 4.8 | $5.00 | $25.00 |
| Claude Sonnet 5 | $3.00 | $15.00 |
| Claude Haiku 4.5 | $1.00 | $5.00 |
| OpenAI | ||
| GPT-5.6 Sol | $5.00 | $30.00 |
| GPT-5.6 Terra | $2.50 | $15.00 |
| GPT-5.6 Luna | $1.00 | $6.00 |
| GPT-5.4 mini | $0.75 | $4.50 |
| GPT-5.4 nano | $0.20 | $1.25 |
| Gemini 3.1 Pro | $2.00 | $12.00 |
| Gemini 3.5 Flash | $1.50 | $9.00 |
| Gemini 3 Flash | $0.50 | $3.00 |
| Gemini 3.1 Flash-Lite | $0.25 | $1.50 |
Sources: OpenAI API pricing · Anthropic (Claude) pricing · Google Gemini API pricing