Anthropic launches Opus 5, halving prices to target enterprise AI demand
Anthropic has released its Claude Opus 5 model at half the cost of its predecessor, aiming to capture enterprise demand for artificial intelligence that delivers clear returns on investment.
Anthropic on Friday launched Claude Opus 5, pricing the new artificial intelligence model at $5 per million input tokens and $25 per million output tokens. The pricing represents a 50 percent reduction compared to its previous flagship commercial model, Fable 5.
Opus 5 outperforms Fable 5 on coding and knowledge work evaluations, capabilities Anthropic said make it "designed to be used every day." The company deliberately avoided positioning the new model as a frontier research tool, stating it remains behind the earlier Mythos 5 release in offensive cybersecurity and biology research.
This targeted capability set and aggressive pricing reflect a critical pivot in the AI sector. Developers are realizing that winning enterprise contracts now requires demonstrating clear returns on investment rather than simply pushing the boundaries of technical performance.
Corporate buyers have largely stopped funding expensive AI experimentation without immediate productivity gains. "Enterprises, in our feedback and with our customer base, are looking for value," said Dianne Penn, Anthropic's head of product management for research. "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."
Anthropic and its primary rival, OpenAI, are both adjusting to this cost-conscious environment. The pricing pressure is intense, coming from established tech giants like Microsoft, Amazon, and Google, alongside a wave of low-cost Chinese startups.
Anthropic's product segmentation also highlights the growing regulatory risks associated with advanced AI. In April, the company released Claude Mythos Preview, which excelled at discovering and exploiting software vulnerabilities. When Anthropic released updated versions of Mythos 5 and the general-use Fable 5 in June, it was forced to temporarily disable both just days later.
The suspension resulted from a government export control directive citing national security authorities. Anthropic restored access to the models after roughly two weeks of negotiations.
Penn noted that Anthropic continues to collaborate with government agencies on pre-deployment testing for Opus 5. By deliberately limiting the new model's "risky, dual-use capabilities," Anthropic is attempting to secure enterprise revenue while insulating its core commercial products from the regulatory bottlenecks that have delayed its most advanced releases.
