Nine major UK banks have just been offered access to OpenAI’s cybersecurity model GPT-5.5 Cyber after they did not gain access to Anthropic’s Claude Mythos models’ latest preview.

The release of Claude Mythos sparked international curiosity and concern from world leaders and organisations, who scrambled to get a look at the yet to be publicly released model.

The model boasts enhanced cybersecurity capabilities, particularly in vulnerability detection, making it a major risk if found in the wrong hands.

To stay competitive, OpenAI released its own cyber-specific model, GPT-5.5 Cyber – The AI Security Institue says that both AI tools show “a similar level of performance” in simulated tasks upon testing.

Upon its announcement, Anthropic released Claude Mythos to 42 companies and organisations, most of which included US tech firms.

Major banks were left out of the release, which meant they could not use the AI tool to test and detect their own systems for vulnerabilities.

Recommended reading

• Claude Source Code Leak Reveals Anthropic’s Secret Plans
• Anthropic Investigates Claimed Unauthorised Access to Mythos
• Anthropic’s Mythos Lands It a White House Meeting
• Is Mythos as Dangerous as Anthropic Claims?

OpenAI, however, has offered access to GPT-5.5 Cyber for UK banks, as well as financial organisations in the EU, Japan, and Canada.

UK banks that now have access to 5.5 Cyber include Lloyds Banking Group, Nationwide, and HSBS.

Santander and NatWest had access to the model through existing agreements with OpenAI.

Anthropic is reportedly looking to make access to its Claude Mythos model public in weeks, sparking concern from the wider cybersecurity community as fears that AI technology could end up in the wrong hands mount.