# Ai-agents
Claude Fable 5.1: The Scary Part Isn't the Benchmark
Anthropic shipped Claude Fable 5.1. It found a crash a team couldn't explain in five years and ran 38 hours unattended. Here's what stood out — and why the model isn't the whole story.
Voice AI in Production: Inside the Vapi x Anthropic Webinar
Notes from the Vapi x Anthropic webinar on building production-grade voice agents with Claude — architecture, guardrails, prompting, and real-world deployments.