How We Built a Token Counter That Supports GPT-5.5 and Claude Opus 4.7

If you work with LLM APIs, you know the pain: you write a prompt, send it, and get a surprise bill because you didn't realize how many tokens it consumed. Our Token Counter solves this by estimating token count and cost before you hit send.

The Challenge: Multilingual Tokenization

English text averages about 4 characters per token in most BPE tokenizers. But Chinese, Japanese, and Korean characters are encoded differently — each character typically consumes 1-2 tokens. Our counter detects character ranges (CJK Unified Ideographs, Hangul, Hiragana/Katakana) and applies appropriate multipliers.

Model Pricing Database

We maintain pricing for 8 models at launch:

• GPT-5.5 — $5/$30 per 1M tokens (input/output)
• Claude Opus 4.7 — $5/$25
• Claude Sonnet 4.6 — $3/$15
• Gemini 3.5 Flash — $1.50/$9
• Gemini 3.1 Flash-Lite — $0.25/$1.50
• Gemini 3.1 Pro — $2/$12
• DeepSeek V4 Pro — $0.435/$0.87
• DeepSeek V4 Flash — $0.14/$0.28

Custom Models

Not every team uses the same models. Our counter lets you add custom models with your own pricing — useful for fine-tuned models, self-hosted endpoints, or newer models we haven't added yet.

Try It

Head to the Token Counter and paste your next prompt to see the breakdown.