Anthropic releases Opus 5 with ‘close’ to Fable 5’s capabilities - The Verge

What the report says

The Verge reports that Anthropic released Claude Opus 5 on Thursday, positioning the new AI model as a high-end option for enterprise users and everyday knowledge work. According to the company’s release, Opus 5 approaches the performance of Claude Fable 5 across many areas and improves on complex coding tasks. Anthropic is also presenting it as suitable for biology-related work, while saying Fable 5 remains the stronger choice for the most demanding projects and long-running AI agents.

The launch follows recent scrutiny of advanced AI systems by the US government. The Verge notes that Claude Fable 5, a public-facing Mythos-class model, had previously drawn government concern, was temporarily taken offline alongside Mythos 5, and later returned with added cyber protections. Anthropic spokesperson Danielle Ghiglieri told The Verge that the company continues to work with government partners on independent model testing, including for Opus 5.

Anthropic emphasized that Opus 5 includes stronger cybersecurity safeguards than Opus 4.8 and described it as the most aligned Opus release so far, with reduced vulnerability to misuse attempts. The Verge places the release in a wider regulatory and industry context, citing OpenAI’s GPT-5.6 rollout in a limited preview to government-approved entities after the Fable 5 episode and a separate OpenAI security incident.

The company is keeping Opus 5 pricing at $5 per million input tokens and $25 per million output tokens, matching Opus 4.8 and making it half the price of Fable 5, according to The Verge. Anthropic is also introducing a research-preview “Fast mode” at double the price, plus an option for customers to use lower-tier fallback models when Opus 5 refuses requests for safety reasons.