What is Coinbase trying to do with Anthropic’s AI?

Coinbase is reportedly seeking access to Anthropic’s Mythos AI model to improve its cybersecurity systems. The model is designed to detect vulnerabilities and defend against advanced cyber threats. This reflects growing pressure on exchanges to strengthen security.

What is Anthropic’s Mythos AI model?

Mythos is a restricted AI model with advanced cybersecurity capabilities. It can identify zero-day vulnerabilities and system weaknesses. Access is limited to select organisations through Anthropic’s Project Glasswing.

What is Project Glasswing and why does it matter?

Project Glasswing is Anthropic’s cybersecurity initiative launched in 2026. It provides partners with tools to detect and fix critical software vulnerabilities. The program highlights how AI is becoming central to security infrastructure.

Why are crypto exchanges focusing more on cybersecurity now?

Crypto platforms are increasingly targeted by sophisticated cyberattacks. AI tools can now automate exploits, making attacks faster and more complex. This raises the stakes for exchanges handling large amounts of user funds.

How does AI make cyber threats more dangerous?

AI can identify and exploit vulnerabilities at scale and speed. Anthropic research showed AI agents could simulate attacks and generate large losses. This makes traditional security methods less effective.

How does Coinbase’s past breach influence this move?

Coinbase experienced a major insider breach in 2025 affecting about 70,000 users. The incident exposed personal data and highlighted internal security risks. This likely increased urgency to upgrade its defence systems.

Who else has access to Anthropic’s Mythos AI?

Major firms like Amazon Web Services, Apple, Google, JPMorgan, Microsoft, and Palo Alto Networks are involved. These are key players in global technology and finance. Their participation shows the importance of AI in cybersecurity.

How could this affect Coinbase’s platform?

If successful, Coinbase could improve threat detection and system resilience. This may reduce the risk of hacks or data breaches. It could also strengthen user trust in the platform.

Image for illustrative purposes only. Not a real photo.

Coinbase eyes Anthropic AI to boost security

Written by Heidi Cuthbert Published Apr 15, 2026, 12:00 AMLast Updated May 7, 2026, 3:07 AM

Coinbase is reportedly in talks with Anthropic to access its restricted Mythos AI model, aimed at strengthening cybersecurity capabilities.

The effort follows Anthropic’s Project Glasswing initiative, which provides select partners with tools to identify vulnerabilities, including zero-day exploits discovered during testing.

The move comes as crypto platforms face increasing threats from AI-powered attacks, with Anthropic research showing autonomous agents can exploit smart contract flaws.

Coinbase’s interest is also shaped by its 2025 insider breach that exposed data from around 70,000 users, highlighting ongoing security challenges.

Anthropic’s Mythos model has been used to uncover long-standing vulnerabilities in critical systems, reinforcing its value for high-security environments.

Access to the model remains limited to select organisations, including major tech and financial firms, with broader availability not planned.

The development underscores intensifying competition to secure digital infrastructure as AI capabilities reshape both offensive and defensive cybersecurity strategies.