GPT-4 Token Counter Online — Free, No Signup, No OpenAI Account

OpenAI charges per token, and counting them before you send a prompt saves money and surprises. For exact counts in production code, use tiktoken. For one-off checks — sizing prompts, estimating cost, fitting documents in the context window — a browser-based counter is faster.

Count GPT-4 Tokens in 5 Seconds

1. Open the Token Counter
2. Paste your prompt or document
3. See instant token count + estimated cost across all GPT models

The counter handles GPT-3.5, GPT-4, GPT-4o, GPT-4o mini, GPT-4.1, GPT-4.1 mini, GPT-4.1 nano, and the o-series reasoning models. No OpenAI account needed.

GPT Pricing by Token (April 2026)

Model	Input ($/M)	Output ($/M)	Context window
GPT-4.1 nano	$0.10	$0.40	128K
GPT-4o mini	$0.15	$0.60	128K
GPT-4.1 mini	$0.40	$1.60	128K
GPT-4.1	$2.00	$8.00	128K
GPT-4o	$2.50	$10.00	128K
o3 mini	$1.10	$4.40	128K
o4 mini	$1.10	$4.40	128K
o3	$10.00	$40.00	128K

The price spread between the cheapest and most expensive OpenAI model is 100x. For most prompts, GPT-4o mini is good enough at 1/17th the cost of GPT-4o.

Common Prompts and Their Token Counts

Prompt type	Approximate tokens	Cost on GPT-4o	Cost on GPT-4o mini
Short question	15	$0.00004	$0.000002
ChatGPT-style chat (with history)	800	$0.002	$0.00012
1-page article to summarize	350	$0.00088	$0.000053
10-page report to summarize	3,500	$0.0088	$0.00053
Code file to refactor	2,500	$0.0063	$0.00038
Long context (full novel)	100,000	$0.25	$0.015

Even an entire novel is 25 cents on GPT-4o or 1.5 cents on GPT-4o mini. The cost of GPT API calls is genuinely tiny for most personal use.

tiktoken vs Browser Counter — Which to Use

Use tiktoken (Python or JavaScript library) when:

• You're building production code that calls OpenAI APIs
• You need exact counts to truncate prompts to a token limit
• You're billing customers based on exact token usage

Use the browser counter when:

• You're checking a one-off prompt
• You want to estimate a project's cost
• You're sizing a document for context window fit
• You don't write Python or JavaScript
• The text contains confidential content you don't want to upload

Cutting GPT-4 Token Usage

If your token counts are higher than budget, the fastest reductions:

• Trim system prompt. Most system prompts have 30-50% filler. Cut redundant instructions and examples.
• Truncate chat history. Keep only the last 5-10 messages, summarize the rest into one short note.
• Use abbreviations and structured format. "User: foo\nAssistant: bar" is shorter than full sentences.
• Set max_tokens for responses. Cap output at the length you actually need.
• Strip markdown if not needed. Bullet points and headers add tokens you may not need.

Test each change with the Token Counter to see exactly how many tokens you saved.