CC Capabilities

Capabilities / Benchmarks

Introducing Claude Opus 4

Anthropic Software engineering score 72/100 confidence 0.88
Category
Benchmarks
Capability
Autonomous multi-agent operation
Observed
2026-05-28
Thesis section
Appendix III, section one: model and benchmark capability evidence

Claim

Claude Opus 4 (claude-opus-4-8) introduces extended thinking, interleaved reasoning, and the ability to run hundreds of parallel subagents unattended in fully autonomous agentic workflows. Anthropic highlights use cases where the model replaces attorneys and engineers, writes full codebases autonomously, and handles open-ended multi-step tasks without human supervision.

Oracle verdict

Anthropic is describing a frontier or production capability that pushes directly on the thesis. The important signal is not the marketing language; it is the widening set of tasks now being routed through model-driven execution rather than ordinary software or headcount.

Why it matters

Manually added from the official Anthropic release stream; observed 2026-05-28. Key signals: hundreds of parallel subagents running unattended, autonomous agentic operation, replacing professional knowledge-worker roles.