Anthropic on Friday announced Claude Opus 5, its newest artificial intelligence model, which the company said matches or exceeds the performance of its Fable 5 model on several industry benchmarks while costing half as much. The launch intensifies an already fierce competition among AI labs that are racing to deliver capable models at prices enterprises are willing to pay.
Opus 5 will cost users $5 per million input tokens and $25 per million output tokens, the same pricing as its predecessor, Opus 4.8, and half the price of Fable 5. Anthropic said the model outperforms Fable 5 on coding and knowledge work evaluations, and that it is designed to be used every day. The company is marketing Opus 5 to its enterprise customers as the ideal everyday model, while positioning Fable 5 as better suited for the most ambitious work and long-horizon AI agent projects.
Dianne Penn, Anthropic’s head of product management for research, told CNBC that enterprises are looking for value. “If it’s a cheaper model or a cheaper offering, but it’s not accomplishing a similar level of quality, it’s actually not useful,” Penn said in an interview.
The release comes at a fraught moment for the industry. Anthropic and its chief rival, OpenAI, have been working to appease a more cost-conscious customer base, where businesses are less inclined to experiment with pricey models without a clear picture of their return on investment. Anthropic also faces pressure from competitors touting cheaper offerings, including Microsoft, Amazon, Google, and a number of Chinese startups. Opus 5 is slightly cheaper than OpenAI’s GPT-5.6, according to The Verge.
Government scrutiny shapes the launch
The Opus 5 release follows a turbulent stretch for Anthropic. In April, the company released Claude Mythos Preview to a select group of users, describing it as extremely capable in cybersecurity and adept at discovering and exploiting software vulnerabilities. In June, Anthropic released Mythos 5 and Fable 5, the latter being a public-facing Mythos-class model that the company said exceeded the capabilities of all other models it had released publicly.
Days after that June release, Anthropic had to briefly disable access to both Mythos 5 and Fable 5 to comply with an export control directive from the government that cited national security authorities. Access was restored after roughly two weeks of intense negotiations. The Verge reported that the Fable 5 concerns and the ensuing negotiations kicked off a new era of AI regulation for the Trump administration, and that when OpenAI released GPT-5.6, it underwent a roughly two-week limited preview solely to government-approved entities.
Anthropic said it gave Opus 5 more cyber safeguards than the previous Opus model. The company described Opus 5 as “the most aligned Opus model and is the least susceptible to being tricked into misuse.” When asked whether Anthropic ran Opus 5 by the Trump administration before the public rollout, spokesperson Danielle Ghiglieri told The Verge, “We continue to work with our government partners to conduct their own independent testing of our models. This includes Opus 5.” Penn said the company is continuing to collaborate with government agencies as part of its pre-deployment and testing.
Anthropic said Opus 5 is not the state-of-the-art for risky, dual-use capabilities like cybersecurity, and that it remains behind Mythos 5 in offensive cybersecurity and biology research. TechCrunch reported that Opus 5 safeguards prevent it from being used to scan for vulnerabilities in a software binary, though it is permitted to search for vulnerabilities in source code, since that task is more likely to serve defensive purposes.
Fewer restrictions and new tools
Opus 5 is free from many of the restrictions that have dogged Fable 5 since its release. Unlike Fable and Mythos, Opus 5 is not subject to the 30-day data retention policy that had raised concerns among privacy-conscious users. Anthropic expects its safety classifiers to engage 85% less often for Opus 5 than for Fable 5, a reflection of the lighter restrictions on the less capable model.
The company also introduced a beta feature called Automatic Fallbacks, which routes requests to a less powerful model when a prompt triggers the safety classifier, giving API users a functional response instead of an error message. Customers can additionally choose a Fast mode for Opus 5, launching in research preview, which offers higher speeds for double the price. The Verge noted that the release subtly addresses recent controversies over allegedly hidden usage caps, which have led to at least one lawsuit, and broader concerns that AI labs are passing costs onto customers in new ways.
Opus 5 arrives only two months after Opus 4.8, which became available on May 28. With Mythos 5, Fable 5, and Sonnet 5 all launching in June, only the lightweight Haiku model remains waiting for an upgrade to the 5 series. Anthropic emphasized that Opus 5 was much stronger at verifying its work and iterating carefully until it succeeds, citing benchmark testing in which the model wrote its own computer vision pipeline in response to an incomplete prompt.