Anthropic introduced Claude Opus 4.7 on April 16, highlighting improvements in software engineering tasks and hard-coding performance. The update is designed to test new safety measures before the full-scale deployment of the more advanced Mythos model to a broader customer base.

The company plans to grant UK financial institutions access to Mythos in the coming week, expanding its limited release beyond select Silicon Valley and Wall Street partners. This move aims to gather real-world feedback on the model’s robustness and practical applications in high-stakes environments.

German banking groups, together with national regulators, have initiated an examination of Mythos to assess its potential to facilitate cyberattacks. The review underscores concerns about large language models’ dual-use capabilities and seeks to establish guidelines for safe deployment in critical sectors.