Opus 5 costs a third of the price — and that’s actually the problem

Anthropic launched Opus 5, positioning it as a much cheaper model with stronger coding performance than its flagship tier.

The company claims state-of-the-art results on several coding and agent benchmarks, and added beta automatic fallbacks plus zero-retention handling.

Stronger guidance for long-running autonomous work could accelerate adoption of agentic coding and broader software automation.

At the same time, teams may need to rethink security, token spend controls, runtime isolation, and microVM-based deployment patterns.