OpenAI offers UK banks cyber tool access as Anthropic limits Mythos – report 

OpenAI has made its GPT-5.5 Cyber system available to nine leading UK banks, while competitor Anthropic has not included them in early access to its Claude Mythos model, reported BBC.  

The two systems are built to detect concealed vulnerabilities in digital infrastructure and can exceed human performance in some cyber security and hacking-related work. 

The move comes after several Japanese financial institutions were given access to OpenAI’s GPT-5.5 model for cyber defence.  

Claude Mythos attracted attention when Anthropic, which recently filed for IPO confidentially, introduced it in April, saying it had identified a flaw in an older system that had gone unnoticed for almost three decades. 

Since then, a number of finance ministers, central bank officials and industry figures have voiced concern that the model could threaten the security of financial systems more broadly. 

The AI Security Institute, which examined both Anthropic’s Mythos and OpenAI’s GPT-5.5 Cyber, said in a report that the two achieved “a similar level of performance” in the tests it carried out. 

Last week, Bank of England Governor Andrew Bailey said UK lenders still could not use Mythos to assess the security of their own systems and applications. 

Bailey asked the US technology company to present the capabilities of its Claude Mythos Preview model to members of the Financial Stability Board (FSB), which he chairs. 

George Osborne, the former UK chancellor who now holds a senior role at OpenAI, told the BBC that Bailey had not approached him directly.  

He said the company did not want to “hide [5.5 Cyber] away or keep it to ourselves”, while also making clear that access would remain limited, as with Mythos. 

“The key things with these tools is that they need to be in the hands of the right people,” he said. 

“We want to make sure that the forces that are establishing order in our democracies have these tools, and the forces that want to disrupt us or commit crime, do not.” 

The UK banks due to receive access to OpenAI’s GPT-5.5 Cyber include Lloyds Banking Group, HSBC and Nationwide.  

NatWest and Santander already use it under existing arrangements. 

A company representative said Anthropic is working quickly to widen access to Mythos.  

The company’s view is that Mythos operates at a higher capability level than 5.5 Cyber, requiring a more cautious approach. 

Anthropic first provided access to a group of 42 companies, most of them US technology businesses. 

OpenAI has allowed use of GPT-5.5 Cyber in the EU, as well as by banks in Japan and Canada, among others. 

Both groups charge users for access, although Anthropic committed $100m to Mythos preview access.