Anthropic releases Opus 5.5 with lower prices and new safety limits

In short: Anthropic has released Opus 5.5, a new version of its Claude AI model, with lower usage prices, faster operation, and updated safety controls.

What happened

Anthropic released Opus 5.5 on Tuesday. The company says it is its strongest-performing model it has tested so far, especially for coding and “knowledge work” tasks, which means work like reading, summarizing, and answering questions.

Anthropic also says Opus 5.5 beats a larger competing model called Fable on many tests, and can complete some informal tasks that Fable could not. In simple terms, it is like getting results that match or exceed a bigger engine, even though the system is smaller or easier to run.

The company lowered prices compared with the previous Opus model. Anthropic says it will charge $20 per million output tokens, down from $25. Tokens are small chunks of text (think of them like the “word pieces” an AI counts when it writes).

Anthropic said Opus 5.5 is less likely to use jargon and more likely to put key information at the start of its replies. The launch comes about two months after Opus 5, and Anthropic said Sonnet 5.5 and Haiku 5.5 are expected in the coming weeks.

Why it matters

Lower prices and faster models can make AI tools cheaper to run for companies and easier to offer to customers. At the same time, Anthropic says Opus 5.5 has strong biology and cybersecurity abilities, so it is released with safeguards that limit harmful uses, like helping find software weaknesses or supporting work related to biological weapons.

Source: TechCrunch AI