Free · private · no sign-up
AI token & API cost calculator
Paste a prompt to count its tokens, then see exactly what it costs on 14 models from 5 providers — per request, per day and per month.
Prices last updated:
Cost on GPT-5.6 Terra
per request
- Input (1,000 tokens)
- $0.002
- Output (500 tokens)
- $0.006
- Per day
- $8.00
- Per month
- $240.00
Token count uses this model’s tokenizer (o200k_base).
Compare every model
Updates live with your inputs. The selected model is highlighted. Click a model for its full price page, or see head-to-head comparisons.
| Model | Provider | Input / 1M | Output / 1M | Context | Per request | Per month |
|---|---|---|---|---|---|---|
| DeepSeek V4 Flash | DeepSeek | $0.14 | $0.28 | 1M | $0.00028 | $8.40 |
| GPT-5.6 Luna | OpenAI | $0.20 | $1.20 | 1.05M | $0.0008 | $24.00 |
| DeepSeek V4 Pro | DeepSeek | $0.435 | $0.87 | 1M | $0.00087 | $26.10 |
| Gemini 3.8 Flash | $0.75 | $3.75 | 1.05M | $0.002625 | $78.75 | |
| Claude Haiku 4.5 | Anthropic | $1.00 | $5.00 | 200K | $0.0035 | $105.00 |
| Grok 4.7 | xAI | $2.00 | $6.00 | 500K | $0.005 | $150.00 |
| Claude Sonnet 5 | Anthropic | $2.00 | $10.00 | 1M | $0.007 | $210.00 |
| GPT-5.6 Terra | OpenAI | $2.00 | $12.00 | 1.05M | $0.008 | $240.00 |
| Gemini 3.1 Pro | $2.00 | $12.00 | 1.05M | $0.008 | $240.00 | |
| Claude Opus 5.5 | Anthropic | $4.00 | $20.00 | 1M | $0.014 | $420.00 |
| Claude Opus 5 | Anthropic | $5.00 | $25.00 | 1M | $0.0175 | $525.00 |
| GPT-5.6 Sol | OpenAI | $5.00 | $30.00 | 1.05M | $0.02 | $600.00 |
| GPT-6 Astra | OpenAI | $10.00 | $50.00 | 1.05M | $0.035 | $1,050.00 |
| Claude Fable 5.1 | Anthropic | $10.00 | $50.00 | 1M | $0.035 | $1,050.00 |
Budget planner
Enter a monthly budget to see how many requests it buys on each model, using the request size from the calculator above.
Request size: 1,000 input + 500 output tokens.
Cheapest model that fits
DeepSeek V4 Flash
$8.40/month for 30,000 requests
| Model | Requests / month | Requests / day | Tokens / month | Fits need? |
|---|---|---|---|---|
| DeepSeek V4 Flash | 357,142 | 11,904 | 535,713,000 | Yes |
| GPT-5.6 Luna | 125,000 | 4,166 | 187,500,000 | Yes |
| DeepSeek V4 Pro | 114,942 | 3,831 | 172,413,000 | Yes |
| Gemini 3.8 Flash | 38,095 | 1,269 | 57,142,500 | Yes |
| Claude Haiku 4.5 | 28,571 | 952 | 42,856,500 | No |
| Grok 4.7 | 20,000 | 666 | 30,000,000 | No |
| Claude Sonnet 5 | 14,285 | 476 | 21,427,500 | No |
| GPT-5.6 Terra | 12,500 | 416 | 18,750,000 | No |
| Gemini 3.1 Pro | 12,500 | 416 | 18,750,000 | No |
| Claude Opus 5.5 | 7,142 | 238 | 10,713,000 | No |
| Claude Opus 5 | 5,714 | 190 | 8,571,000 | No |
| GPT-5.6 Sol | 5,000 | 166 | 7,500,000 | No |
| GPT-6 Astra | 2,857 | 95 | 4,285,500 | No |
| Claude Fable 5.1 | 2,857 | 95 | 4,285,500 | No |
Understand what you’re paying for
- What is a token?A plain-English explanation of AI tokens: how tokenizers split text, how many tokens a word or page uses, and why token counts differ between models.
- How AI API pricing worksHow AI providers bill for API usage: per-million-token rates, input vs output prices, context windows, batch discounts, caching and long-context surcharges.
- Input vs output token cost, explainedWhy output tokens cost 2–6× more than input tokens, how to estimate the split for your workload, and practical ways to keep output costs down.
- How to estimate AI API costs before you buildA step-by-step method to forecast monthly AI API spend before writing code: measure prompts, model the output, multiply by volume and add a safety margin.
- Token cost by use case: chatbot, RAG and summarizationWorked token and cost estimates for three common AI workloads — a support chatbot, retrieval-augmented generation (RAG) and document summarization.
How this calculator works
Token counts for OpenAI models use the o200k_base tokenizer, loaded into your browser only when you start typing. Other providers don’t publish a browser tokenizer for their current models, so we estimate their counts from the same text with a per-provider adjustment and label them as estimates. Read what a token is for why counts differ.
Costs use each provider’s standard list price per million tokens from our public pricing file. Cached-input discounts, priority tiers and long-context surcharges aren’t applied — see each model page for those notes.