Benchmarks
Measured results on document comparison: how accurately AI assistants and comparison tools find the changes between two versions, with the method and data behind each figure.
“When Claude is asked to compare two versions of a contract and has no comparison tool, it writes one.”
BenchmarkFrom 
What It Costs Claude to Compare Two Documents Without a Comparison Tool
5.7 minutes and 858,000 tokens per comparison for Claude. 2.8 seconds and none for the API.
“The agent spends most of its tokens getting to the changes, not reasoning about them.”
BenchmarkFrom 
Markdown, Word or PDF: What a Redline Costs an AI Agent to Read
An AI agent reads a Markdown redline with 44% fewer tokens than Word and 69% fewer than PDF.
