Prompt Cost Calculator
Paste your prompt, set output length, and see exactly what it costs across all 34 AI models. No signup required.
Your Prompt
Type or paste a prompt above to see costs across all models.
How Token Counting Works
This tool estimates tokens using the standard ~4 characters per token ratio for English text. Actual token counts vary by model's tokenizer — GPT models use ~4 chars/token, Claude uses ~3.5, and Gemini uses ~4. For production-accurate counts, use your provider's official tokenizer. Our estimates are typically within 10-15% of actual counts, which is sufficient for cost estimation and model comparison.
Understanding the Costs
Prices shown are per 1M tokens. Input costs are based on your prompt length. Output costs depend on the response length (prompt length × output multiplier). The "Savings" column shows how much you save compared to the most expensive model for your workload. Streaming mode adds 15% to output costs to account for SSE framing and repeated context tokens in streamed responses. For most use cases, budget models like Gemini Flash, DeepSeek, or Mistral Small offer 90%+ savings versus premium models with minimal quality loss for routine tasks.
Related Tools
- Token Estimator — Count tokens and estimate costs
- Cost Calculator — Estimate monthly costs for any model
- Cost Explorer — See all 34 models ranked by cost
- Cheapest AI API Finder — Find the cheapest model
- Monthly Spend Estimator — Compare costs across all 34 models