Anthropic Releases 3.5 Haiku for Claude Users
Date: 2024-12-13 Author: Oliver Abernathy Categories: BUSINESS
The development of 3.5 Haiku was announced in October along with another version, 3.5 Sonnet. However, the actual launch of the model took place later. In terms of power, 3.5 Haiku corresponds to the company's previous flagship neural network, Claude 3 Opus.
As Anthropic emphasized, 3.5 Haiku is focused on such tasks as programming, working with data, including its extraction and labeling, as well as content moderation. Its key advantage is the ability to generate longer text responses compared to its predecessor, Claude 3.
In addition, the neural network was trained on the basis of fresh data, which allows it to take into account recent events and provide more relevant information. However, 3.5 Haiku does not support image analysis, making it less versatile than the 3.5 Sonnet model.
In November, Anthropic CEO Dario Amodei proposed requiring AI companies to conduct mandatory security testing before releasing new products.
Anthropic has previously made its technology available to US intelligence and defense agencies as part of its strategy to collaborate with government agencies.
