Anthropic releases Sonnet 5.5 AI model with faster responses

In short: Anthropic has released Sonnet 5.5, a new version of its Sonnet AI model that the company says is faster and cheaper to run.

What happened

Anthropic announced Sonnet 5.5, the latest version of its mid-range AI model in the Claude family. An AI model is the software “brain” behind a chatbot, like the engine inside a car.

Anthropic says Sonnet 5.5 is aimed at everyday work tasks, including writing office documents and helping with coding. The company says it is about 30 percent faster than Sonnet 5, which came out about three months ago.

Anthropic also says Sonnet 5.5 uses fewer tokens. Tokens are small chunks of text that AI systems count, a bit like billing by the word. Using fewer tokens can lower costs for people and businesses that pay based on usage.

Anthropic positions Sonnet as less powerful than its top model, called Opus, but sometimes more practical because it can be quicker and cheaper. The company says its own tests show Sonnet 5.5 doing better than Opus 5.5 on some “agent” coding tasks. In this context, an agent is like a helper bot that can take steps on its own, such as writing and checking pieces of code.

Anthropic also said Sonnet 5.5 has “comparable” cyber capabilities to an older Opus model. Because of that, Anthropic says Sonnet 5.5 will get the same cyber safeguards as some of its higher risk models. Anthropic also plans to release a new version of Haiku, its smallest model, in the coming weeks.

Why it matters

Faster and cheaper AI tools can make it easier for more workplaces to use AI for writing and coding, but added safety controls matter when a system could also be used for harmful cyber activity.

Source: TechCrunch AI