Anthropic is releasing the AI model Claude Opus 5, which follows its predecessor Opus 4.8. It is intended as a workhorse for everyday use, but in many respects it rivals Claude's high-performance model Mythos 5.
In its press release, Anthropic particularly highlights Opus 5's more efficient operation: Specifically, Opus 5 is said to deliver significantly more performance than Opus 4.8 at the same cost, especially in software development. Anthropic refers to various internally tested benchmarks here.
Anthropic's current models and competitor model ChatGPT 5.6 in Anthropic's internal benchmark comparison.
(Image: Anthropic)
In the benchmark table, it is particularly noticeable that Opus 5 often comes close to or even surpasses Anthropic's high-end model Fable 5, even though using Fable 5 costs twice as much, according to the Claude documentation.
