Developer tools

Context budget

Reserve output space using explicit token counts.

Processed locally

This does not count text. Enter model-specific counts and check its output limit. Enter zero overhead if it is already in the input count.

Your answer will appear here.

Help and method

Add input, reserved output and additional overhead. Subtract this allocation from the context limit; report excess separately. Whole counts: 0–10¹²; context must be positive.

Token counting and scopeMethods and reliability

Worked example

An example you can try or adapt to your own values.

Context limit in tokens
8000
Counted input tokens
5000
Reserved output tokens
2000
Additional overhead tokens
500

Result

8,000 context − 5,000 input − 2,000 output − 500 overhead → 500 tokens remaining.