The context window is the maximum number of tokens an LLM can process in a single request. It includes your prompt, system instructions, and the model's response.
Context Window Sizes
- GPT-4: 8K-128K tokens
- Claude: Up to 200K tokens
- Gemini: Up to 1M tokens
- Open-source models: Varies widely
Context Window Considerations
- Larger context = higher cost
- Performance may degrade with very long contexts
- "Lost in the middle" problem
- RAG can work around limits
How many tokens is 1 word?
Roughly 1.3 tokens per word in English. A 4K token context is about 3,000 words.
Practical example
A document workflow combines system instructions, user text, retrieved passages, tool results, and output allowance.
What to check
Count with the relevant tokenizer and inspect truncation, rejected requests, and task quality as context grows.
Limits and tradeoffs
A larger context limit does not mean the model reliably uses every included fact.
Maintained by DriftRail. AI assisted the October 2026 content and template revision. Examples are illustrative; they are not measured product benchmarks. Suggest a correction.