Anthropic has updated its Claude frontier model to integrate computer usability, and it seeks to improve the performance of tasks that require “hundreds of steps” to complete.
The family of artificial intelligence (AI) models has expanded this Tuesday with the new Claude 3.5 Haiku, an updated version of Claude 3 Haiku, Anthropic’s fastest model, which offers low latency, better following of instructions and a more precise use of tools.
“It matches the performance of the Claude 3 Opus, our largest previous model, in many reviews for the same cost and similar speed to the previous generation Haiku,” the company detailed on its official blog.
Anthropic has also updated Claude 3.5 Sonnet, which improves its performance in programming and the use of agent tools, and is also the first AI model to offer the new computer use capacity in public beta.
This capability means developers can ask the model to use computers the same way a person does: “looking at a screen, moving a cursor, clicking buttons, and typing text.”
By teaching Claude general computer skills, the company paves the way for him to use a wide variety of tools and software programs designed for people. It also makes it easier to perform repetitive tasks that require dozens, and sometimes even hundreds, of steps to complete.