Anthropic has released Claude Opus 5, a new large language model currently leading the Artificial Analysis leaderboard ahead of the competing Claude Fable 5. The company describes it as a thoughtful and proactive system that approaches frontier intelligence levels while costing half the price of the previous generation. Pricing matches Opus 4.8, and a fast mode option remains available at twice the standard rate.
The update demonstrates a specific capability where the model generated its own computer vision pipeline to reconstruct a 3D machine part from a static image without direct visual access. Security testing shows the system identifies vulnerabilities effectively but lacks training on how to exploit them into active cyber threats. This distinction suggests the model prioritises defensive analysis over offensive capabilities.
* Opus 5 is priced the same as the prior Opus 4.8 version
* The model performs better at finding security flaws than at using them
* Anthropic has published a dedicated prompting guide for the release
