Anthropic unveils Claude Opus 5: near-Fable 5 performance at half the price — cyber safety measures eased, with automatic fallback on refusal

Anthropic launched Claude Opus 5, the successor to Opus 4.8, claiming near-Claude Fable 5 performance at roughly half the price.

The model is now available on Claude Pro, Claude Max, Claude Code, and the Claude API, with new effort controls and a research-preview fast mode.

Anthropic also updated cyber-task safety handling and added automatic fallback behavior, signaling a more flexible approach to model operations.

The release raises the bar on frontier model price-performance and could influence enterprise adoption, developer workflows, and AI safety debates.