I tested Opus 4.8 against 4.7 using coding, medical, finance, and legal traps, then cross-checked the results with multiple AIs.

Source: ZDNet — AI — read the full report at the original publisher.

This is a curated wire item. The Continuum Brief does not republish full third-party articles; this entry links to the original source.