English version available

This tool compares advertised LLM context windows against their real-world reliable zones, based on needle-in-haystack benchmarks. Enter your document length to see which models can reliably process it.