Anthropic rolls out Claude Opus 4.7, an AI model that is 'broadly less capable' than Mythos

On Thursday, Anthropic introduced its latest artificial intelligence model, Claude Opus 4.7. While the company touts this version as an enhancement over its predecessors, it has also acknowledged that it falls short in capabilities compared to the recently launched Claude Mythos Preview, which stands as Anthropic's most advanced AI model. Claude Mythos Preview has garnered attention for its proficiency in identifying vulnerabilities and security risks within software, and it is currently being rolled out to a limited number of companies under a new cybersecurity initiative known as Project Glasswing. Despite the advancements of Claude Opus 4.7, Anthropic admits that it does not reach the performance levels of its Mythos counterpart. However, the new model does demonstrate notable improvements over its earlier version, Claude Opus 4.6, which debuted in February. Anthropic emphasized that with Opus 4.7, they are implementing enhanced safeguards designed to automatically recognize and block requests associated with prohibited or high-risk cybersecurity activities. The insights gained from deploying these safeguards in real-world scenarios are expected to inform the company’s future efforts, ultimately paving the way for a broader release of Mythos-level models.