Anthropic just launched Claude Opus 5, its latest model, which it says boasts near Fable 5 capabilities at a fraction of the cost.
“Claude Opus 5 is available today. It’s a thoughtful and proactive model that comes close to the frontier intelligence of Claude Fable 5 at half the price,” the AI giant shared in a press release. “Opus 5 is designed to be used every day: it works more efficiently than other models.”
Anthropic says the model performed great on coding and knowledge work evaluations, though Anthropic’s secretive top dog model Mythos 5 still leads on cybersecurity tasks. That’s on purpose, according to Anthropic itself.
“As with its predecessor, Opus 4.8, we’ve intentionally avoided training Opus 5 on cyber tasks,” the company said. “The model has nevertheless improved substantially on these tasks as a result of becoming more generally capable, and it comes close to Mythos 5 at finding cybersecurity vulnerabilities.”
But the model won’t exploit those vulnerabilities as much as Mythos 5 might. In the release, Anthropic also claimed that Opus 5 was its “safest model yet,” and that it exhibits “the lowest rates of deceptive behavior; and is the least susceptible to being tricked into misuse.”
Anthropic’s Fable 5 was supposed to be the consumer-friendly version of Mythos, which has only been made available to a select group of organizations and government agencies after its cyber capabilities were deemed far too advanced and dangerous. But only a couple of days after it debuted to the public, Fable 5 itself was banned by the Trump administration as well, after Amazon executives reportedly notified the government of a jailbreak in the model’s cybersecurity safeguards. After negotiations with the government and what Anthropic has described as stronger safeguards, the model was re-released to the public by the end of June.
Scarred once before, Anthropic now seems to be really focusing on the cyber safety guardrails of its new product release. The Opus 5 has also received independent testing by government partners, per The Verge.
Anthropic’s new cost-efficient alternative comes at a time when the rising cost of AI compute and a tokenmaxxing trend have prompted American businesses to rethink their AI budgets. Frontier AI models might be great at certain tasks, but that performance comes at a hefty price. One example is Uber: the company went all in on AI adoption, only to race through its entire 2026 AI budget within the first four months of the year. With the returns of that investment not materializing just yet, Uber COO Andrew Macdonald said in May that the AI investment was becoming “harder to justify.”
Because companies prize their bottom line and might still believe in immense productivity gains from AI, reports suggest businesses have started experimenting with Chinese AI models instead, which are generally much cheaper than their American alternatives. Until recently, the assumption was that American models might be pricier, but their capabilities were still months ahead of what any Chinese offering can provide. A new model release from Chinese startup Moonshot challenged that notion last week.
According to independent testing, Moonshot’s Kimi K3 outperforms many of its American alternatives, including matching or beating Anthropic’s Fable 5 on some benchmarks. Even though the open model is costlier than other Chinese offerings, taken at face value, it is still cheaper than many American models, which sent Silicon Valley into a bit of a spiral over the past week, so much so that the Trump administration felt the need to step in and tell people to calm down.
Read the full article here
