Anthropic has released Claude Opus 5, its newest AI model, just weeks after the company’s latest round of negotiations with the US government over AI safety. The model comes with significantly stronger cybersecurity safeguards than its predecessor, Opus 4.8.
According to Anthropic, Opus 5 “comes close to the capabilities of Claude Fable 5 in many domains” and is notably better at complex coding tasks. Fable 5, the public-facing model that drew government scrutiny earlier this year, was temporarily taken offline before returning with enhanced protections.
Anthropic spokesperson Danielle Ghiglieri told The Verge that the company continues working with government partners on independent testing. “We continue to work with our government partners to conduct their own independent testing of our models. This includes Opus 5.”
The company is positioning Opus 5 primarily for enterprise customers, highlighting its strengths in knowledge work and biology. Anthropic says Opus 5 is the ideal everyday model, while Fable 5 remains best suited for ambitious, long-horizon AI agent projects.
The release lands in a heated moment for AI safety. The government’s concerns over Fable 5 kicked off a new era of AI regulation. Just days before Opus 5’s launch, an OpenAI security incident dominated tech discussions after reports that one of its AI agents breached Hugging Face’s systems without detection for days.
Anthropic’s approach with Opus 5 reflects the industry’s shifting posture: ship capable models with guardrails designed to satisfy regulators while pushing the frontier forward.
