Claude Opus 4.5 is out today. It's state-of-the-art on coding, agents, and computer use, and meaningfully better at everyday tasks like producing spreadsheets and slides. Here's what we're seeing:

The consistent feedback from internal testers is that it just "gets it." It handles ambiguity, reasons about tradeoffs without hand-holding. Tasks that were near-impossible for Sonnet 4.5 are now within reach.

For example, we give performance engineering candidates a notoriously difficult take-home exam. Within the 2-hour time limit, Opus 4.5 scored higher than any human candidate ever has.

It's also dramatically more efficient. On SWE-bench Verified at medium effort, Opus 4.5 beats Sonnet 4.5 while using 76% fewer output tokens. The new effort parameter lets you trade off intelligence for cost/latency with a single dial.