The news: OpenAI is granting several additional UK banks access to its GPT-5.5 cybersecurity model, while rival Anthropic continues to limit access to its more powerful Claude Mythos tool, per BBC. OpenAI’s move comes as UK banking leaders have expressed frustration over their inability to test Mythos, which Anthropic has kept restricted due to concerns about potential misuse.

Zoom in: The divide between OpenAI and Anthropic reflects two competing strategies for earning banks' trust. Anthropic has tightly restricted access to Claude Mythos, arguing that its advanced vulnerability-discovery capabilities require heightened safeguards. OpenAI has taken a more expansive approach, extending GPT-5.5 access to additional financial institutions (FIs) while emphasizing responsible use. 

The contrast underscores a broader issue facing banks: whether the greatest value lies in gaining early access to powerful AI tools, or partnering with providers that prioritize tighter controls and risk management.

Why that matters: The competition is about whose model performs better in addition to how banks gain access. As banks increasingly adopt AI for areas ranging from employee productivity and software development to risk management and cybersecurity, they may also get ahead of competitors by gaining access to emerging AI tools. OpenAI is leveraging broader enterprise partnerships to introduce new AI capabilities to FIs. 

Santander and NatWest had preexisting relationships with OpenAI and already have access to GPT-5.5, while the AI platform will expand access to Lloyds, HSBC, and Nationwide. Meanwhile, Anthropic has limited Mythos access to a smaller group of vetted organizations while it evaluates the model's risks. 

The companies’ differing strategies suggest trust will depend on how each balances early deployment and hands-on testing with controls that limit potential misuse. As cyber AI becomes more integrated into security operations, those initial access decisions could influence which providers become longer-term strategic partners.

Implications for banks: The development could force banks to rethink how they evaluate AI vendors. As advanced cyber AI becomes increasingly embedded in vulnerability management, software development, and security operations, FIs may place greater weight on governance, access policies, and risk controls. Over time, decisions about which AI providers to partner with may influence broader technology strategies and potentially shape which AI ecosystems banks build around in the years ahead.

This content is part of EMARKETER’s subscription Briefings, where we pair daily updates with data and analysis from forecasts and research reports. Our Briefings prepare you to start your day informed, to provide critical insights in an important meeting, and to understand the context of what’s happening in your industry. Non-clients can click here to get a demo of our full platform and coverage.