Home/Answers/Input and output tokens
AI COST LAB / Input and output tokens

How to Choose Input and output tokens

Choose Input and output tokens by use case and compatibility first, specs second; test locally when possible.

Updated 2026-11-01About 3 minLocal test
Quick answer

Choose Input and output tokens by use case and compatibility first, specs second; test locally when possible. Input includes prompt/context/tool results; output includes answer/reasoning; their ratio drives cost.

Why it matters

Stacked specs do not guarantee fit for your use case.

Long context adds input.

Step by step

  1. Measure input, output and retry tokens per user/task.
  2. Separate cache, batch, embeddings and fine-tuning costs.
  3. Use 30 days of real requests for cost per user, then margin.
  4. Model prices change; baseline from official pricing pages and record the date.

Common mistakes

  • Estimating from input tokens only; output often costs more.
  • Forgetting retries, failures and tool calls.
  • Assuming cache hits by default.
  • Using stale prices for long-term margin.

Checklist

  • Input/output/cache/retry tokens are tracked separately.
  • Batch/async fit and latency requirements are clear.
  • Embeddings/fine-tuning/image/audio are separate.
  • Cost per user and margin use real usage.
  • Price date and fallback model are documented.

Key facts

  • Long context adds input.
  • Long answers add output.
  • Track separately.

FAQ

How do I choose Input and output tokens?

Choose Input and output tokens by use case and compatibility first, specs second; test locally when possible. Input includes prompt/context/tool results; output includes answer/reasoning; their ratio drives cost.

How long does it take?

Usually 1-3 minutes; for intermittent issues, run 30 seconds and retest elsewhere.

Is data uploaded?

Runs locally and does not upload video/audio; extensions may still read the page.

Source: Anthropic