Prompts · Client-side only
Token & context counter
Estimate token counts for the five major LLM families and see what percentage of each context window your prompt will consume.
100 chars16 words
| Model family | Est. tokens | Context | % used |
|---|---|---|---|
| Claude (Anthropic) | 27 | 200,000 | 0.01% |
| GPT-4o / GPT-5 (OpenAI) | 25 | 400,000 | 0.01% |
| Gemini 2.5 (Google) | 25 | 1,000,000 | 0.00% |
| Llama 3/4 (Meta) | 28 | 128,000 | 0.02% |
| DeepSeek V3 | 26 | 128,000 | 0.02% |
Token counts are approximations based on family-average characters-per-token ratios for English. Actual counts vary with language, punctuation, code, and JSON. For exact counts, use each provider's tokenizer.