Opus 4.8

Anthropic's most capable model (released May 28, 2026; knowledge cutoff Jan 2026). Same $5/$25 price as 4.7/4.6. **Headline gain is reliability/honesty: ~4× less likely than Opus 4.7 to let code flaws pass unremarked, lower rates of deception/misaligned behavior, better at flagging its own uncertainty.** Best for: production-ready code, sophisticated agents, complex document creation, long-running tasks, creative writing, architectural decisions, strategic advising, emotional support. Adaptive thinking, defaults to high-effort. Strongest computer-use/browser-agent model Anthropic has tested