FrontierThe story, in brief
Anthropic's Claude Opus 4.7 makes a big leap in coding, while deliberately scaling back cyber capabilities
Claude Opus 4.7 just reset the coding benchmark. Anthropic deliberately weakened its hacking skills to get there.
The KeyNews take
Why it matters
Anthropic's new flagship model shows measurable coding improvements while the company takes an intentional stance on capability trade-offs—signaling a shift in how leading labs balance performance against safety risks.
The key facts
4 to knowClaude Opus 4.7 flagship release with major coding improvements
Anthropic deliberately reduced cybersecurity/hacking capabilities during training
Capability trade-off strategy: performance gains paired with intentional safety constraints
Published April 16, 2026
Go to the source
The Decoderthe-decoder.com
Publisher excerpt: Anthropic's new flagship model Claude Opus 4.7 delivers major improvements in coding tasks. During training, the company deliberately tried to reduce certain cybersecurity capabilities.