Anthropic Unveils Sonnet 5.5, a Faster, Cheaper AI Model After CEO’s Call to Slow Innovation

Anthropic, an AI company known for its cautious approach, has launched a new model called Sonnet 5.5. This comes just weeks after its CEO, Dario Amodei, urged the industry to slow down rapid AI development to better manage safety risks.

Sonnet 5.5 is designed as a more affordable and quicker option compared to Anthropic’s higher-end models. It’s aimed at everyday work like coding, fixing bugs, and creating documents or presentations. While it doesn’t push the limits of AI capabilities, it improves on tasks compared to its previous version, Sonnet 5.

One of the key benefits is efficiency. The company says Sonnet 5.5 runs more than 30% faster and costs up to 30% less per task because it uses fewer computing resources. It also includes stronger cybersecurity protections, similar to those found in Anthropic’s top-tier models, which helps prevent misuse and safeguards sensitive information.

Sonnet 5.5 follows Opus 5.5, a more powerful and costly model Anthropic released earlier this month. Both belong to the company’s new Claude 5.5 family of AI models. Sonnet 5.5 is priced at $2 per million input tokens and $10 per million output tokens, making it a budget-friendly choice for users who don’t need the full power of the flagship models.

Anthropic CEO Amodei has been vocal about the need to slow AI’s rapid advance. In early September, he proposed giving independent experts permanent access to AI systems to check safety and alignment with human values. This cautious stance comes amid concerns about AI’s potential risks, including misuse and unintended consequences.

Overall, Sonnet 5.5 offers a practical, safer AI tool for many routine tasks, reflecting Anthropic’s balanced approach of innovation combined with careful risk management.