AI Token Estimator
Paste your prompt and instantly estimate token counts and API costs across Claude, GPT-5, and Gemini. Runs entirely in your browser — your text never leaves your device.
Pricing disclaimer: Cost estimates use approximate published rates and may be outdated. Always verify current pricing on the provider's official pricing page before making budget decisions.
How Token Estimation Works
Large language models process text as tokens, not characters or words. A token is roughly 3–4 characters of English text, a common word, or a punctuation mark. Models are billed per token consumed, so estimating tokens before making API calls helps you control costs and fit within context window limits.
This tool uses a BPE-style (Byte Pair Encoding) heuristic that estimates tokens from character and word boundaries. For typical English prose it stays within 5–10% of the actual tokenizer count. Code, special characters, and non-Latin scripts may have higher variance.
Counting the tokens is the easy half. Bringing the number down is the other half, and our guide to token optimization in production works through the levers that actually move a monthly invoice. The gap between a single-prompt estimate and a real bill is widest inside agent loops, where every tool result is re-sent on the next turn — we measured that directly when a web-search agent spent most of its tokens on context it never read.
Context Window vs. Max Output
Each model has a maximum context window — the total tokens it can process in one request (input + output combined). If your prompt alone is large, you may have limited room left for the model's response. The progress bars above show what fraction of each model's context window your input text would consume.
Model Comparison
| Model | Context | Input / 1M | Output / 1M | Best for |
|---|---|---|---|---|
| Claude Sonnet 4.6 | 200K | ~$3.00 | ~$15.00 | Complex tasks, coding, analysis |
| Claude Haiku 4.5 | 200K | ~$1.00 | ~$12.00 | Fast, affordable, high-volume |
| GPT-5.5 | 1M | ~$12.00 | ~$30.00 | Multimodal, broad compatibility |
| GPT-5.4 mini | 1M | ~$0.75 | ~$4.50 | Low-cost, high-throughput |
| Gemini 3.1 Pro | 1M | ~$2.00 | ~$12.00 | Very long documents, RAG |
| Gemini 3.5 Flash | 1M | ~$1.50 | ~$9.00 | Budget-friendly, large context |
Prices are approximate and subject to change. Verify at the provider's official pricing page before making cost decisions.
Related reading
Adding One Tool to Your Agent Wiped the Whole Prompt Cache
Effloow Lab ran 17 OpenAI API calls. Appending, deleting, reordering or rewording a single tool zeroed the prompt cache every time. One setting avoided it.
Read →GPT-5.6 Prompt Caching: The 24-Hour Cache Is Gone
GPT-5.6 added explicit prompt cache controls and quietly took the 24-hour cache away. We ran the API and measured what that swap costs.
Read →GPT-5.6 Programmatic Tool Calling: When It Cuts the Token Bill
We ran GPT-5.6's programmatic tool calling on a real API. On one task it cut tokens 92%. On another it cost 2.2x more. Here's the rule that decides which.
Read →Find these tools useful?
Get one short weekly dispatch with new tools, guides, and what we tested. No spam, unsubscribe anytime.