Anthropic has released Claude Opus 5, a new version of its flagship AI model that the company pitches less as a leap in raw power and more as a leap in efficiency.
According to The Star, Anthropic released the model on July 24 and says it approaches the performance of the company's most powerful system at half the price. Social Samosa similarly reports that Anthropic is claiming better performance at lower cost, while InfoWorld frames the release around a "more efficient" model and The Hindu describes it as an efficiency upgrade.
On the feature side, a TechJuice explainer reports that Opus 5 ships with "thinking" turned on by default — meaning the model works through problems step by step without being asked — along with a 1 million-token context window, enough to hold very large volumes of text in a single conversation. TechJuice characterizes the combination as frontier-level intelligence at lower cost.
Verdict reports that the model is available across all of Anthropic's platforms, rather than being limited to a single product or preview tier.
The common thread across coverage is price-performance rather than benchmark bragging. For most of the past few years, each new frontier model arrived with a higher bill attached, which kept the most capable systems out of reach for routine, high-volume work. A model that gets close to top-tier quality at half the cost changes that math for developers and businesses deciding what they can afford to automate.
It also signals where competition in AI is heading: not just who has the smartest model, but who can deliver near-frontier intelligence cheaply enough to run it everywhere.
Why it matters: when frontier-grade AI gets meaningfully cheaper, it stops being a showcase feature and starts being infrastructure that ordinary software can afford to build on.