Claude Opus 5 Launches with Enhanced Performance and Cost Efficiency

Claude Opus 5 is now available and is positioned as a cost-effective alternative to Claude Fable 5, offering advanced capabilities at half the price.

The model is recognized as the new standard for coding and knowledge work evaluations, outperforming its predecessor, Opus 4.8, in various benchmarks, particularly in software engineering tasks.

Opus 5 shows significant improvements in life sciences evaluations, especially in organic chemistry and protein-related tasks, scoring higher than Opus 4.8.

In terms of alignment and safety, Opus 5 is noted for being the most aligned model to date, with lower rates of deceptive behavior compared to previous models. It maintains strong safeguards against risky capabilities, particularly in cybersecurity and biology.

While it performs well in identifying vulnerabilities, Opus 5 does not advance in exploiting those vulnerabilities compared to Mythos 5. The model's cybersecurity classifiers are less restrictive than those on Fable 5, allowing for some vulnerability detection while blocking more dangerous tasks.

Opus 5 is priced at $5 per million input tokens and $25 per million output tokens, with a Fast mode available for increased processing speed. It is accessible on all platforms and through the Claude API.