Context Window Calculator

A context window represents the maximum total tokens (input + output) an LLM can process in a single API call. Context limits range from 128,000 tokens (GPT-4o) up to 1,048,576 tokens (Gemini 3.5 Flash and Claude Sonnet 4.6), and even 10,000,000 tokens (Llama 4 Scout).

Context Window Capacity Calculator

Input Unit

Number of Words

Standard English Estimate: 100,000 tokens (actual counts vary by model tokenizer below)Approximate Characters: 400,000 characters

Context Window Consumption by Model

Visual representation of how much context window is consumed based on your input parameters.

Llama 4 Scout (17B)97,403 / 10,000,000 (0.97%)

Gemini 3.5 Flash100,000 / 1,048,576 (9.54%)

Claude Fable 5102,740 / 1,000,000 (10.27%)

Claude Sonnet 4.6102,740 / 1,000,000 (10.27%)

GPT-5.596,154 / 1,050,000 (9.16%)

GPT-4o96,154 / 128,000 (75.12%)

GPT-4Exceeded (1205%)

o4-mini96,154 / 200,000 (48.08%)

Qwen 3.6 27B98,684 / 262,144 (37.64%)