Claude Opus 5.5: Anthropic's new flagship gets 40% cheaper

TVS News AI, editorially supervised, presents Claude Opus 5.5: Anthropic's new flagship gets 40% cheaper.

Anthropic has unveiled Claude Opus 5.5. It may sound like routine model maintenance at first – but it's exciting for anyone using AI tools at work, in daily life, or in their own projects. Because the new flagship is not only set to become cheaper, but also safer, according to the maker.

According to Golem.de, the new model performs on par with the company's previous top-tier models – while costing 40 percent less, according to Anthropic. For anyone already experimenting with Claude or using it in production, this is a statement of intent.

We summarize what has emerged from the official announcement so far – and where Anthropic still owes us the details.

What is Claude Opus 5.5?

Anthropic is promoting the new model in the official announcement as its "showpiece model." The Golem.de report indicates that the AI system is expected to perform on the level of an earlier flagship model. Exactly how the comparison shakes out – which predecessor serves as the benchmark and across which benchmarks – cannot be clearly read from the short version of the announcement so far.

What does Claude Opus 5.5 cost?

The biggest concrete number from the announcement: 40 percent less. What that means in euros per token, per API call, or compared to competing models like GPT or Gemini doesn't come out of the current reporting. What's clear is only this: Anthropic itself is talking about a price cut of this size versus comparable predecessors. Once the final price list is out, it will be easier to gauge what this means for developers and businesses.

What makes Opus 5.5 safer?

Anthropic leads its announcement title with the keyword "safer." What concrete improvements that covers – whether in handling sensitive data, in robustness against prompt injection, or in new safety tests – doesn't come out of the current coverage. Here too, it's a question of waiting to see what details Anthropic delivers in the coming days.

What does this mean for you?

If you're already using Claude or thinking about getting started, it's worth a look at the new pricing structure. Once Anthropic publishes the final API prices and security features, you can better assess whether switching to Opus 5.5 pays off for you – and how the model compares to GPT-5 or Gemini 2.5. Until then: wait, compare prices, then decide.