ai-tokens

ai-tokens 💰

Stop guessing. Know exactly what your AI API calls cost—before you make them.

Real-time token counter and cost calculator for GPT-5, Claude 4.5, Gemini 2.5, DeepSeek, and 20+ AI models. Built for developers who are tired of surprise API bills.

😱 The Problem

# You write a prompt...
$ cat my-prompt.txt | your-ai-tool

# Later that month...
💸 Your OpenAI Bill: $347.23

What happened? You had no idea that your 100-line context was costing $2.50 per call.

✨ The Solution

# Check BEFORE you send
$ ai-tokens count my-prompt.txt

━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
           TOKEN ANALYSIS
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━

Model: GPT-4o

Input Tokens:    2,847 tokens
Est. Cost:       $0.01424

Output (est):    500 tokens
Est. Cost:       $0.00750

Total Cost:      $0.02174

💡 Cost Optimization:
  • Switch to gpt-4o-mini → Save 96% ($0.0004)
  • Reduce context by 1k tokens → Save $0.005

━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━

Now you know. $0.02 per call. Make an informed decision.

🚀 Quick Start

# No install required
npx ai-tokens count "Your prompt here"

# Or install globally
npm install -g ai-tokens

💡 Features

✅ Accurate Token Counting

Uses official tokenizers (tiktoken for OpenAI, estimations for others)

✅ Real-Time Cost Estimates

Up-to-date pricing for GPT-5, Claude 4.5, Gemini 2.5, DeepSeek, and 20+ models

✅ Model Comparisons

See exactly how much you'd save with cheaper alternatives

✅ Works with Files

ai-tokens count ./prompts/analyze-code.txt

✅ Works with Pipes

cat large-context.md | ai-tokens count -m claude-opus-4

✅ Beautiful Output

Color-coded warnings, optimization tips, comparison tables

📖 Usage Examples

Count tokens in a file

ai-tokens count prompt.txt
# Shows: tokens, cost, optimizations

Specify model

ai-tokens count prompt.txt --model claude-sonnet-4
ai-tokens count prompt.txt -m gpt-4o-mini

Set expected output length

ai-tokens count prompt.txt --output 1000
# Estimates cost for 1000 token response

Compare across models

ai-tokens compare prompt.txt

# Output:
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
       MODEL COST COMPARISON
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━

Current: GPT-4o ($0.0097)

┌──────────────────┬──────────┬────────┬─────┐
│ Model            │ Cost     │ Savings│ %   │
├──────────────────┼──────────┼────────┼─────┤
│ Gemini Flash     │ $0.0002  │ +$0.01 │ 98% │
│ GPT-4o Mini      │ $0.0004  │ +$0.01 │ 96% │
│ Claude Haiku     │ $0.0030  │ +$0.01 │ 70% │
└──────────────────┴──────────┴────────┴─────┘

🎯 Best Alternative: gemini-1.5-flash
   Save $0.0095 (98%)

List all models

ai-tokens models

# Shows pricing for all supported models:
# - GPT-4o, GPT-4, GPT-3.5
# - Claude Opus, Sonnet, Haiku
# - Gemini Pro, Flash
# - Llama 3.1 (405B, 70B, 8B)

Simple output (for scripts)

ai-tokens count prompt.txt --simple
# Output: 2847 tokens | $0.0217 | gpt-4o

🎯 Real-World Examples

Example 1: Code Review Prompt

$ ai-tokens count code-review-prompt.txt

Input Tokens: 3,245 tokens
Total Cost:   $0.0487

💡 This costs more than a coffee ☕
   Consider using gpt-4o-mini to save 95%

Example 2: Documentation Generation

$ ai-tokens compare docs-prompt.txt -m gpt-4

Current Model: GPT-4
Cost: $0.157

🎯 Switch to GPT-4o → Save $0.141 (90%)

Example 3: Bulk Processing

$ for file in prompts/*.txt; do
    ai-tokens count "$file" -m claude-sonnet-4 --simple
  done

# Quick cost estimate for all prompts

🔥 Why This Tool Exists

Stop getting surprised by your AI bill.

• ✅ Know costs BEFORE calling the API
• ✅ Compare models instantly
• ✅ Optimize prompts for cost
• ✅ Track spending across projects
• ✅ No more "$500?! WTF?!" moments

📊 Supported Models

Latest Models (2025)

Model	Input Price	Output Price
GPT-5	$1.25/M	$5.00/M
GPT-5 Mini	$0.15/M	$0.60/M
Claude Sonnet 4.5	$3.00/M	$15.00/M
Gemini 2.5 Pro	$1.25/M	$5.00/M
Gemini 2.5 Flash	$0.30/M	$1.20/M
DeepSeek R1	$0.14/M	$0.28/M

Popular Models

Model	Input Price	Output Price
GPT-4o	$2.50/M	$10.00/M
GPT-4o Mini	$0.15/M	$0.60/M
Claude Opus 4	$15.00/M	$75.00/M
Claude Sonnet 4	$3.00/M	$15.00/M
Claude Haiku 3.5	$1.00/M	$5.00/M
Gemini 2.0 Flash	FREE	FREE
Gemini 1.5 Pro	$1.25/M	$5.00/M
DeepSeek Chat	$0.14/M	$0.28/M

20+ models supported - Pricing auto-updated from LiteLLM community data

Prices as of January 2025

🛠️ How It Works

• Tokenization: Uses tiktoken (OpenAI's official tokenizer) for GPT models
• Estimation: Smart approximations for Claude, Gemini, Llama
• Pricing: Real-time pricing data for all major providers
• Display: Beautiful, color-coded CLI output

📦 Installation

# Global install
npm install -g ai-tokens

# Or use directly
npx ai-tokens count "your prompt"

# Or add to project
npm install ai-tokens

🎨 CLI Output Features

• 🎨 Color-coded costs - Green = cheap, Yellow = moderate, Red = expensive
• 📊 Comparison tables - Side-by-side model pricing
• 💡 Smart suggestions - Automatic optimization tips
• ⚠️ Cost warnings - Alerts for expensive prompts
• 🎯 Best alternatives - Instant cheaper model recommendations

🔮 Coming Soon

• Budget tracking (ai-tokens budget set 50)
• Interactive mode (ai-tokens interactive)
• Markdown reports (ai-tokens report)
• Git hooks (block expensive commits)
• VS Code extension
• Browser extension (for ChatGPT UI)

🤝 Contributing

Pull requests welcome! Especially for:

• Additional model support
• Improved token estimations
• More cost optimization suggestions

📄 License

MIT © Daniel Shashko

👤 Author

Daniel Shashko

• GitHub: @danishashko
• LinkedIn: daniel-shashko

💬 Share Your Savings

Found this useful? Tweet your cost savings!

Stop guessing. Start knowing. 💰