Token Cost Calculator
Work out what a workload costs across models before you commit to one. Runs in your browser, sends nothing anywhere.
These token counts are estimates
This is an estimate based on characters per token, not a real tokenizer. Every model family tokenizes differently: Anthropic notes that Claude 4.7 and later produce roughly 30% more tokens for the same text than their earlier models. Expect to be within about 10 to 20% for English prose, and further out for code, non-Latin scripts, and heavily structured text. For exact counts, use your provider’s token counting API.
| Model↕ | Provider↕ | In $/M↕ | Out $/M↕ | Per call↕ | Per day↕ | Per month▲ |
|---|---|---|---|---|---|---|
| Gemini 2.5 Flash-Lite | 0.10 | 0.40 | $0.0018 | $0.180 | $5.40 | |
| GPT-5.4 nano | OpenAI | 0.20 | 1.25 | $0.0045 | $0.450 | $13.50 |
| Gemini 3.5 Flash-Lite | 0.30 | 2.50 | $0.0080 | $0.800 | $24.00 | |
| GPT-5.4 mini | OpenAI | 0.75 | 4.50 | $0.017 | $1.65 | $49.50 |
| Claude Haiku 4.5 | Anthropic | 1.00 | 5.00 | $0.020 | $2.00 | $60.00 |
| GPT-5.6 Luna | OpenAI | 1.00 | 6.00 | $0.022 | $2.20 | $66.00 |
| Gemini 3.6 Flash | 1.50 | 7.50 | $0.030 | $3.00 | $90.00 | |
| Claude Sonnet 5Introductory pricing through 31 August 2026. Rises to $3 / $15 on 1 September 2026. | Anthropic | 2.00 | 10.00 | $0.040 | $4.00 | $120.00 |
| Gemini 3.1 Pro PreviewRate shown is for requests up to 200k tokens. Above that it is $4 input and $18 output. | 2.00 | 12.00 | $0.044 | $4.40 | $132.00 | |
| GPT-5.6 Terra | OpenAI | 2.50 | 15.00 | $0.055 | $5.50 | $165.00 |
| GPT-5.4 | OpenAI | 2.50 | 15.00 | $0.055 | $5.50 | $165.00 |
| Claude Sonnet 4.6 | Anthropic | 3.00 | 15.00 | $0.060 | $6.00 | $180.00 |
| Claude Opus 5 | Anthropic | 5.00 | 25.00 | $0.100 | $10.00 | $300.00 |
| Claude Opus 4.5 | Anthropic | 5.00 | 25.00 | $0.100 | $10.00 | $300.00 |
| GPT-5.6 Sol | OpenAI | 5.00 | 30.00 | $0.110 | $11.00 | $330.00 |
| GPT-5.5 | OpenAI | 5.00 | 30.00 | $0.110 | $11.00 | $330.00 |
Click any column to sort. Monthly figures assume 30 days. Rates are standard API pricing: batch processing and prompt caching both cost less, so treat these as an upper bound.
Prices checked on 2026-07-26. Model pricing changes often. Confirm with the provider before you budget against these numbers.