Anthropic’s Opus 5 bets cheaper, safer AI will matter more than a breakthrough

Anthropic has launched its latest AI model, Claude Opus 5, which it claims offers performance comparable to its most powerful model, Fable, at half the price. The release is part of a trend toward more frequent, iterative updates focused on cost-efficiency and performance rather than major capability leaps.
Anthropic’s Opus 5 bets cheaper, safer AI will matter more than a breakthrough

Anthropic’s Opus 5 bets cheaper, safer AI will matter more than a breakthrough
Anthropic’s latest launch says a lot about where the AI race is heading. Claude Opus 5 is not being sold as a moonshot; it’s being sold as a sharper bargain — nearly Fable-level performance, Anthropic says, without the Fable-level bill.

That framing is the real story. Anthropic describes Opus 5 as a model that “comes close to the capabilities of Claude Fable 5 in many domains” at roughly half the price. Axios and Ars Technica read the release the same way: not as a dramatic leap in raw capability, but as another sign that top labs are now competing on efficiency, pricing, and product segmentation as much as benchmark glory. The sticker price helps make the case: $5 per million input tokens and $25 per million output tokens, in line with Opus 4.8 and below Fable’s cost.

Anthropic is also trying to thread a political needle. After scrutiny over cyber risk, the company says Opus 5 is “the most aligned Opus model and is the least susceptible to being tricked into misuse,” while The Verge reports it ships with stronger cybersecurity safeguards and ongoing government testing. TechCrunch, though, notes the model is still “cheaper and less restrictive than Fable,” with fewer safety interventions and no 30-day data-retention policy attached to the more tightly controlled top-tier systems.

That leaves a familiar split in the market. Supporters see an everyday enterprise model: capable enough for developers and knowledge workers, but cheaper to run and easier to deploy. Early user reactions are more mixed. Every called the model “brilliant in flashes, frustrating in practice,” saying it argued with instructions and struggled with existing toolchains. And while third-party enthusiasm arrived fast — Perplexity said Opus 5 “outperformed all but Fable 5, while being 57% cheaper” — that praise only reinforces the central point: this release is about value, not awe.

In other words, Opus 5 looks less like a new chapter in AI and more like the industry’s new normal.

Continue reading https://foxvector.com/stories/019fafc1-591e-2cd0-70ad-157ef880b4ac

Write a comment