Anthropic releases Claude Opus 5.5, and it's not just more powerful: it's cheaper, too

Published Sep 22, 2026, 12:44 PM EDT Simon is a Computer Science BSc graduate who has been writing about technology since 2014, and using Windows machines since 3.1. After working for an indie game studio and acting as the family's go-to technician for all computer issues, he found his passion for writing and decided to use his skill set to write about all things tech. Since beginning his writing career, he has written for many different publications such as WorldStart, Listverse, and MakeTechEasier. However, after finding his home at MakeUseOf in February 2019, he would eventually move on to its sister site, XDA, to bring the latest and greatest in Windows, Linux, and DIY electronics. Sign in to your XDA account Summary Claude Opus 5.5 launches, delivering higher performance than Opus 5. Matches Claude Fable 5.1 for most tasks, with notable coding and knowledge gains. Costs drop ~40%: input $4/1M and output $20/1M — big savings for heavy users. In the fast-paced world of AI agents, it's common to see companies release more capable models; however, it's rarer to see them get more intelligent and cheaper to use at the same time. Well, Anthropic has managed to achieve just that, as Claude Opus 5.5 promises to be a blessing to both your workflows and your wallet. Claude Opus 5.5 is out, and the stats are impressive It's both more powerful and more affordable than Opus 5 Credit: Anthropic Over on the ClaudeAI X feed, Anthropic announced the release of its newest model, Opus 5.5. Right off the bat, the company makes two pretty stunning claims: that it performs similarly to its very powerful Claude Fable 5.1 for "most tasks," and that it costs 40% less to run than Opus 5. If true, these claims imply that the new version of Claude isn't just mighty, but also a huge money-saver for power users who tend to chew through credit. The ClaudeAI X post shared a chart showing the performance gains, which you can see in the image earlier in the article. We're seeing a ~11-percentage-point increase in agentic coding capabilities with Terminal-Bench 4.0, a 6-percentage-point increase in CursorBench 4.0 task capabilities, and a higher score for knowledge work. The company also claims that, while Claude Opus 5 costs $5 for 1 million input tokens and $25 for 1 million output tokens, Claude Opus 5.5 slices that down to $4 and $20, respectively. This should be a significant boon for Claude power users, as they now have access to a model that's both stronger and cheaper to run than what they've been using.