Anthropic Just Released Its Latest AI Model. It Says It Doesn’t Advance The Frontier.

Anthropic launched its latest AI model, Claude Sonnet 5.5, on Monday, describing it as a “clear upgrade over” its predecessor, running 30% faster and costing up to 30% less for most work.

“Sonnet 5.5 improves on Sonnet 5 across domains—in some cases dramatically. On several evaluations, Sonnet 5.5 at Max effort even performs comparably to Opus 5.5. However, benchmark scores capture only one facet of a model’s capabilities; in our own testing, and in that of external testers, Opus 5.5 remains clearly stronger at complex, open-ended work requiring sustained judgment,” the company detailed. It also said it is better at coding, completing scoped tasks and creating documents, slides and spreadsheets.

Theo Chu, a research product manager at Anthropic, told CNBC that the product is “really for the cost-conscious customer where they might not need as much intelligence.”

“It might be routine tasks that just need execution, but don’t need that judgment that Opus can bring,” he added.

Anthropic noted that the model doesn’t advance frontier capabilities. The clarification comes as the company and other top actors in the AI industry keep calling for the pacing of AI development over the risk it could get out of control.

On Monday, a group of top figures in the industry published a paper calling on governments to prepare for the possibility of what could be the “most consequential technological development in history.”

The paper noted that AI is “rapidly automating AI research and development,” recalling that in late August Anthropic noted that its systems were completing more than 25% of internal AI R&D work with only high-level supervision. That figure only stood at 1% five months earlier. OpenAI also reported that its systems routinely did such tasks. “Based on benchmark trends and other evidence, some experts think that AI R&D could be fully automated within the next few years,” the document added.

Such a scenario could lead to an “intelligence explosion,” compressing years of AI progress into months or less.” This could then allow frontier development to expand exponentially, with a feedback loop “overcoming frictions such as diminishing returns to research labor and

The intelligence explosion would bring benefits but also dangers. The first category means that improvements could “rapidly obtain superintelligent AI systems that transform many sectors of society, bringing unprecedented breakthroughs in science and industry.”

“However, at the same time they could “could arrive much faster than society can adapt to the resulting risks, such as AI-enabled pandemics, cyber attacks on critical infrastructure, and large-scale labor disruption.”

Consequently, it calls on policymakers to get visibility into AI R&D automation, develop ways to “steer and constrain an intelligence explosion,” and prepare “society for the impacts of an intelligence explosion.”

U.S. policymakers do not appear to intend to slow down AI development at the moment. Attorney General Blanche said in an interview on Sunday that the Trump administration does not need new AI regulation and must ensure that it remains the leading AI power.