Anthropic Says Opus 5 Deliberately Avoids Cybersecurity and High-Risk Training

Anthropic reveals that Claude Opus 5 deliberately avoids training on cybersecurity tasks. Strict evaluations confirm that the model does not push high-risk dual-use capabilities in biological research or offensive cyber operations to new boundaries, still lagging behind Mythos 5.

2026-07-25 ~ 2026-07-25 · 2 related posts

• Anthropic says Claude Opus 5 was intentionally left untrained on cyber tasks — rez0__ · 2026-07-25
• Anthropic says Opus 5 still trails Mythos 5 on exploit generation despite better vulnerability finding — TheZvi · 2026-07-25