Anthropic Launches Opus 5 AI Model at Roughly Half the Price of Fable 5

 AI startup Anthropic on Friday launched Opus 5, an updated Claude model that the company said delivers capabilities approaching those of its flagship Fable 5 system at roughly half the price, according to a report by Reuters.

The San Francisco-based developer stated that the new model was built to handle daily office workflows and computer programming tasks. In an interview with Reuters, Dianne Penn, a product leader at Anthropic, noted that Opus 5 offers higher efficiency than the company’s Opus 4.8 release from May, indicating a rapid development pace across its software generations.

"We're building and continue to consistently deliver frontier intelligence and bring that as accessibly as possible with every model generation," Penn said.

Operational Guidance and Pricing Structure

Anthropic outlined distinct functional roles for its current software lineup. Penn recommended that users select Opus 5 for everyday value and routine work, while reserving Fable 5 for "days-long, very autonomous projects" requiring greater operational independence.

• Claude Opus 5: Tailored for daily office tasks and coding, offering near-frontier performance at roughly half the cost of Fable 5.

• Claude Fable 5: Positioned as Anthropic's top-tier system intended for multi-day autonomous projects.

Safety Testing and Security Evaluations

During internal testing, Opus 5 showed lower performance in cybersecurity capability evaluations compared to Anthropic's flagship Fable 5 model, Reuters reported. Consequently, Anthropic applied less restrictive safety guardrails to Opus 5 than those enforced on Fable 5.

Anthropic also stated that internal testing suggested Opus 5 was less susceptible to prompt manipulation or being tricked into misuse than its other existing software models.

Reuters reported that Anthropic temporarily restricted access to Fable 5 following concerns raised by U.S. officials regarding potential risks that its advanced capabilities could be diverted to foreign military intelligence operations.

Context on Open-Weight Alternatives

When asked about competing software releases, including the Kimi K3 open model developed by Beijing-based Moonshot AI—which U.S. officials accused of extracting capabilities from Anthropic models—Penn addressed the broader landscape of open-weight systems.

Unlike proprietary closed systems, open-weight architectures allow users to download, run, and customize model parameters locally. Penn noted that it "remains to be seen" how open-weight models perform on complex, multi-step real-world projects that users typically assign to Claude.

Reporting & Source Credits

• Primary Source Article: Reuters — Anthropic rolls out Opus 5 AI model in efficiency upgrade (Reporting by Jeffrey Dastin in San Francisco; Editing by Sonali Paul).

• Publication Date: July 24, 2026

• Last Updated: July 25, 2026

• Photo Credit: REUTERS / Dado Ruvic / File Photo