Anthropic says its Claude Opus 5 AI model is an improvement that takes less effort to deliver similar results as the nearly-banned Fable 5.
Claude Opus 5 is available now as the next version of what used to be Anthropic’s highest performer. When Fable 5 was released, recalled, then released again, it sat above Opus 4.8 in terms of pure performance.
The margins have shifted slightly. Claude Opus 5 reportedly costs half the price of Fable 5, yet it delivers results that are similar in performance. Benchmarks released by Anthropic detail the disparity. It looks like Fable 5 is slightly better in a few areas, but the difference is narrow. In areas like business workflows, computer use, and agentic research, Claude Opus 5 performs better.
On several coding and knowledge work evaluations, Opus 5 is the new state-of-the-art: pic.twitter.com/Cl3fDxM0gP
Fable 5 is famously a token-monster, too. Anthropic says it improved the Opus model, all while keeping the costs similar to what users were used to with Opus 4.8. Anthropic also notes that even when Opus 5 runs at minimal effort, it still succeeds in more tasks than other models at high effort when working through business workflows.
Fable 5 is currently available to users, but it carries heavier restrictions. The new Opus 5 release, though, is expected to intervene around 85% less often. Fable 5 is still king in terms of highly advanced tasks in cybersecurity exploitation and biology, but Opus 5 is expected to present a much more capable model to the broader user base.
Opus 5 is available on all platforms starting today.
FTC: We use income earning auto affiliate links. More.
Check out 9to5Google on YouTube for more news:
