In this illustrated photo, the Anthropic logo is displayed on a smartphone, with the Claude mythological logo displayed in the background.
Sopa Images | Light Rocket | Getty Images
human The company is sharing access to its cutting-edge AI model, Mythos, with the EU after the bloc sought access for months over cybersecurity concerns.
The European Commission confirmed to CNBC on Monday that it had “several productive meetings” with U.S. AI companies.
“We welcome the latest developments regarding possible future access,” Thomas Renier, the EU’s technology sovereignty spokesperson, said in an emailed statement, adding that the EU aims to gain a clearer picture of the potential risks posed by the technology.
Anthropic introduced Mythos to a limited number of businesses in April as part of its cybersecurity venture, Project Glasswing. This model is excellent at identifying security flaws and weaknesses in software. The announcement sparked a wave of concern about cybersecurity threats posed by malicious actors.
“Remember, Mythos is not a one-off; a new wave of powerful models is coming to market,” Renier said. “This is a common challenge and we are stepping up consultations with like-minded partners, including the United States.”
Formal terms of the deal have not yet been disclosed. Anthropic did not immediately respond to CNBC’s request for comment.
The EU stepped up talks with the U.S. government last week to gain access after Anthropic told the European Commission it needed government permission, sources told CNBC earlier.
Officials said at the time that the U.S. government was opposed to sharing the powerful model with the entire non-U.S. government because it aims to remain the dominant leader in AI.
The EU had already been granted access to OpenAI’s GPT-5.5-Cyber model in May, but the European Commission said at the time that it had met with Anthropic “four or five times” and that the agreement was at “different stages.”
When Mythos was first released in April, governments, banks, and technology companies were wary of the risks it posed. This technology can uncover thousands of previously unknown software vulnerabilities that could be exploited by malicious actors to accelerate cybercrime.
The Trump administration also announced an agreement with the United States in May. google deep mind, microsoft And Elon Musk’s xAI allows governments to evaluate artificial intelligence models before they are made public.
