Context Window Demo

Total Tokens: 0 Context Window: 100
About This Demo

This demonstration shows how token context windows work in Large Language Models:

  • Each message shows its token count
  • Messages outside the context window appear faded
  • Like an LLM's active memory, only the most recent tokens within the context window are available for processing - older tokens "fade" from memory

Want to learn more about Tokens and how different LLMs Tokenize?
Check out the Tokenizer Playground.