Latest Claude Opus "4.7" released: "Can handle difficult coding," "follows instructions properly without ignoring them" — prompt readjustment recommended

Anthropic has released Claude Opus 4.7 to the public, with stronger coding, image understanding, and instruction following.

The model also reduces cyberattack capability and adds safety features that detect and block risky use cases automatically.

Prompts tuned for earlier models may need to be readjusted.

This update raises both productivity and safety, but may require changes to existing workflows and deployments.