Anthropic PBC is unveiling a new artificial intelligence model designed to handle a wide range of enterprise tasks at a reduced price point. The move comes as the startup prepares for its highly anticipated Wall Street initial public offering and seeks to maintain its competitive edge over rivals.

The company officially launched Claude Opus 5.5 on Tuesday, stating that the model performs comparably to the more powerful Fable 5.1 across most work-related tasks, while its usage costs are substantially lower than the previous Opus 5 version. Anthropic noted that the new model generates output faster than Opus 5 and delivers clearer, more concise writing.

The AI developer plans to roll out new versions of Claude Sonnet and Haiku in the coming weeks, expanding its 5.5 series model family. Anthropic is accelerating efforts to build more advanced and cost-efficient AI software as it competes for customers against OpenAI, Alphabet-owned Google, and numerous Chinese technology firms.

The Claude developer anticipates launching its IPO as early as this autumn, while competitor OpenAI is expected to wait until next year. Opus 5.5 also marks the first new model release since Anthropic Chief Executive Officer Dario Amodei called on the industry to support a slower pace of AI development.

Amodei stated that the company will implement new safety measures, including bringing in third-party evaluators. This follows growing warnings across Silicon Valley and beyond about the potentially catastrophic risks associated with AI. Anthropic says Opus 5.5 was tested by external organizations such as METR and Frontier Design prior to its release.

The company has also deployed security guardrails similar to those used in its more powerful Fable model, designed to limit high-risk applications in cybersecurity and biological research.