BT Joins Anthropic's Project Glasswing as AI Evolves Cybersecurity

WHO:

TechRadar - BT is the first UK firm to join Anthropic's Project Glasswing.

The Guardian - Anthropic is the maker of the Claude artificial intelligence models.

WHAT:

TechRadar - BT declared its support for Anthropic’s Project Glasswing, which uses the Claude Mythos Preview AI model to protect networks against cyberattacks.

The Guardian - Anthropic released a new version of its technology, dubbed Fable 5, to the general public while restricting its use in sensitive areas.

The Guardian - Fable 5 is promoted as useful for writing and debugging software code, answering complex research questions, and analyzing images.

The Guardian - Anthropic expanded its select group of cybersecurity partners in Project Glasswing to about 200 organizations in more than 15 countries.

The Guardian - The launch of Fable 5 comes with a price tag of $10 per million input tokens and $50 per million output tokens.

The Guardian - Anthropic's Opus 4.8 model was made public in late May and is billed as less capable than Fable 5.

WHERE:

The Guardian - Anthropic is based in San Francisco.

WHEN:

The Guardian - Fable 5 was made widely available on a Tuesday.

The Guardian - The select group of cybersecurity partners was expanded in early June.

WHY:

The Guardian - Anthropic restricted access to Mythos on cybersecurity grounds due to its ability to quickly identify vulnerabilities in critical infrastructure, including banking platforms and power grids.

The Guardian - The US government tested the model over security concerns after a legal dispute with Anthropic.

HOW:

The Guardian - Anthropic hired outside experts to spend more than 1,000 hours trying to find ways to bypass restrictions, a process known as red-teaming.

The Guardian - The company ran a bug bounty program to find security flaws, with no one finding a way to completely unlock the model.

(UPDATED June 10, 2026 06:24 UTC)

BT has become the first UK company to join Anthropic's Project Glasswing, which aims to enhance cybersecurity using the Claude Mythos Preview AI model. In related news, Anthropic has released a new AI version called Fable 5, designed for tasks like writing software code and analyzing images, while also expanding its cybersecurity partner network to about 200 organizations across more than 15 countries. The launch of Fable 5 comes with a pricing structure of $10 per million input tokens and $50 per million output tokens, and access to the Claude model has been restricted due to its potential to identify vulnerabilities in critical infrastructure.