Nine UK banks now have access to OpenAI's GPT-5.5 Cyber through the company's TAC program. The move gives these institutions a dedicated tool for cybersecurity applications. Meanwhile, Anthropic has blocked all UK banks from using its Mythos Preview model, creating a sharp contrast in AI provider strategies.
TAC Program Rollout
OpenAI's selective TAC program is the channel for this cybersecurity AI deployment. The nine banks are using GPT-5.5 Cyber specifically for threat detection and response tasks. This isn't a public rollout—access remains limited to these financial institutions under the program's terms. The companies involved haven't indicated plans to expand the list of participating banks.
Anthropic's Unexpected Restriction
Anthropic made a clean break by blocking UK banks from Mythos Preview entirely. The company provided no explanation for the move. This shutdown happened while OpenAI was opening its cybersecurity tool to UK banks. The restriction covers all UK banking institutions without exception.
Limited Options for Banks
UK banks now face diverging paths with these major AI providers. The nine TAC participants have GPT-5.5 Cyber at their disposal for security work. Every UK bank remains shut out of Anthropic's Mythos Preview. There's no indication these banks are seeking alternatives elsewhere. The immediate reality is fewer available AI models from leading companies for this sector.
OpenAI's TAC program continues with its current bank participants. Anthropic maintains its block on UK bank access to Mythos Preview.
