LLM watermarking degrades AI agent performance and speed

Understanding the Impact of LLM Watermarking on AI Agent Behavior

Lasso Security tested LLM watermarking on AI agents and found it hurts tool-calling accuracy by 3–5 points on BFCL V3, adds 11 seconds of latency, and increases output length by 20–30%. The post doesn't name the specific models or agent frameworks tested.

Why it matters: Has concrete benchmark data answering a production-relevant question: what's the performance cost of watermarking on agents. But the post doesn't disclose which models/frameworks were tested, so the numbers are directional only — hence the score cap.