Context Window Calculator
Check input, output reserves and remaining model context locally.
Numeric token budgets · browser-local
Your token budget
Include system messages, conversation history and tool definitions in input. Reserve room for reasoning/thinking and generated output.
COMBINED CONTEXT
Remaining tokens
—
1,050,000 shared input + output tokens · 128,000 maximum output
Enter your input and output budget, then check context.
This model shares its context window between input and reserved output. Approaching limit means at least 80% used; it is a planning signal. Negative remaining tokens mean the window is exceeded.
Limits verified 2026-10-08 · Official model source
Enable JavaScript to calculate locally. No server submission is available.
Leave room for the answer.
A context window limits how much a model can process in one request. Budget the complete conversation, system instructions and tool definitions, plus expected output. The calculator checks the selected model’s published limits without sending anything to a provider.
Work from token counts.
Use actual API usage or the Token Counter’s clearly labeled estimate, then enter the numbers here. Navigation never transfers prompt text or saves your budget.
Understand the limits.
Does output use the same window?
For combined windows, input and reserved output share the limit. Models with separately published input/output limits are checked separately. The selected model’s semantics are shown alongside the result.
What does approaching the limit mean?
At 80% utilization, the tool signals that remaining room is limited. Exactly 100% leaves no headroom. Only usage above the published window is labeled as exceeding it. Output has its own maximum check.
Will the provider accept my request?
This is planning arithmetic, not request validation. Hidden wrappers, tool tokens, thinking budgets, account settings and endpoint behavior can change actual capacity. Check the linked model documentation.