Anthropic Unveils Claude Opus 5, a Cost-Focused Upgrade to Its AI Suite

Anthropic announced the launch of Claude Opus 5, describing it as the company’s “best-performing and most cost-effective” model. The Verge reports the rollout occurred on a Thursday, while CNBC notes an announcement on the following Friday, indicating a brief discrepancy in timing. Opus 5 is positioned as an everyday model for knowledge work and biology, with the company emphasizing stronger cybersecurity safeguards than its predecessor, Opus 4.8, and a reduced susceptibility to misuse.

The Verge

According to both outlets, Opus 5 surpasses the earlier Claude Fable 5 on coding and knowledge-work benchmarks, though Anthropic concedes it is not the state-of-the-art for “risky, dual-use capabilities” such as cybersecurity. The model is marketed toward enterprise customers seeking reliable performance for routine tasks, while Fable 5 remains targeted at more ambitious, long-horizon AI projects. Anthropic also highlights a “Fast mode” that offers higher speeds for double the price and an automatic fallback to lower-tier models when safety safeguards intervene.

The Verge

Opus 5 is priced at $5 per million input tokens and $25 per million output tokens, matching the cost of the prior Opus 4.8 and representing half the price of Claude Fable 5. The Verge notes the price is also slightly lower than OpenAI’s GPT-5.6. The “Fast mode” option incurs double the standard rate, and customers may select a research-preview feature that enables this higher-speed tier.

The Verge

Anthropic spokesperson Danielle Ghiglieri said the company continues to work with government partners for independent testing of Opus 5, underscoring ongoing collaboration with U.S. regulators. Dianne Penn, head of product management for research, told CNBC that enterprises are “looking for value” and that a cheaper model must still deliver comparable quality to be useful.

CNBC

• “Enterprises, in our feedback and with our customer base, are looking for value,” — Dianne Penn

CNBC