Anthropic rolls out Opus 5 AI model in efficiency upgrade

SAN FRANCISCO — Anthropic has launched Opus 5, its latest artificial intelligence model, which the company says approaches the capabilities of its more powerful Fable 5 model at half the price  positioning it as an everyday option for businesses, knowledge workers and developers.

The San Francisco-based lab said the new Claude model is well suited to routine office work and computer programming tasks. Opus 5 is priced at $5 per million input tokens and $25 per million output tokens, the same as its predecessor, Opus 4.8, which was released in May.

In an interview with Reuters, Anthropic product leader Dianne Penn said the release reflected the company's rapid pace of development, adding that Anthropic aimed to deliver frontier intelligence as accessibly as possible with each new generation of models.

The company is also introducing an effort “dial” that lets users decide how much computing power the model devotes to a given task. At lower settings, Anthropic says, Opus 5 can preserve much of its performance while using fewer tokens and costing less to run. A separate new option allows users to switch models mid-task, which can help keep costs down.

According to Anthropic, Opus 5 was less capable of exploiting cyber vulnerabilities than its top-tier model in testing, meaning its related safeguards are less restrictive than those applied to Fable 5. The company said the model was also less susceptible than its other current systems to being manipulated into misuse.

Penn suggested users choose Opus 5 for value and Fable 5 for “days-long, very autonomous projects.” Fable 5, released in June, was temporarily withdrawn earlier this year amid US concerns that its capabilities could be diverted to foreign military intelligence, before access was restored.

The launch is Anthropic's fourth model in the Claude 5 family in under two months, underscoring how competition in the AI industry has shifted from occasional blockbuster launches toward rapid, incremental gains in capability, cost and speed. The company continues to compete with OpenAI and Google for enterprise customers increasingly focused on whether models can handle real workloads affordably.