Skip to main content

วิธีการคำนวณ LLM Context Window

learn.whatIsHeading

An LLM context window calculator shows how much of a model's context limit a given amount of text consumes, and estimates the input cost. Context window is the maximum text a model can process at once.

สูตร

context_tokens_available = model_context_window - system_prompt_tokens - response_tokens_reserved
window
Context window (tokens) — Max tokens the model accepts
system
System prompt (tokens) — Tokens used by system instructions
response
Max response (tokens) — Tokens reserved for output
available
Available for input (tokens) — Tokens left for user input/history

คำแนะนำทีละขั้นตอน

  1. 1Context window measured in tokens (1 token ≈ 0.75 words)
  2. 2Input tokens include both the prompt and any prior conversation
  3. 3Exceeding context limit causes earlier content to be forgotten
  4. 4Cost = (context tokens ÷ 1000) × input price per 1K tokens

ตัวอย่างที่มีคำตอบ

อินพุต
32K tokens used in 128K model
ผลลัพธ์
25% context used, ~$0.08 input cost (GPT-4o)
อินพุต
Full 200K context (Claude)
ผลลัพธ์
~150,000 words, ~600 A4 pages
อินพุต
1,000 token conversation
ผลลัพธ์
~750 words, minimal cost at most price points

คำถามที่พบบ่อย

What is a context window?

The maximum number of tokens a model can process in a single request. Longer contexts = more memory and latency. Claude 3.5 Sonnet: 200K tokens.

How do I estimate tokens in my prompt?

Roughly: 1 word ≈ 0.75 tokens; 1 line of code ≈ 5–10 tokens. Use official tokenizer tools for precision.

What happens if I exceed the context window?

The request fails or tokens are truncated. Always verify your total token count (system + input + expected output).

พร้อมที่จะคำนวณแล้วหรือยัง? ลองใช้เครื่องคิดเลข LLM Context Window ฟรี

ลองด้วยตัวคุณเอง→

การตั้งค่า