Anthropic Unveils 3.5 Haiku for Claude Users

Anthropic has launched a new artificial intelligence model, 3.5 Haiku, for users of the Claude chatbot platform. The neural network is available on both web and mobile devices. 

@AnthropicAI launched Claude 3.5 Haiku. Now available to everyone pic.twitter.com/Jxjp9B5Ww2

— Chubby♨️ (@kimmonismus) December 12, 2024

3.5 Haiku was introduced in October alongside the updated version 3.5 Sonnet, although its launch was anticipated later. Its performance matches the company’s previous flagship neural network, Claude 3 Opus. 

According to Anthropic, 3.5 Haiku is best suited for programming, data extraction and labeling, and content moderation.

The neural network can generate longer text fragments compared to the third version. It is trained on current information, enabling it to reference recent events.

3.5 Haiku does not support image analysis, making it less versatile compared to 3.5 Sonnet.

In November, Anthropic CEO Dario Amodei urged other AI companies to conduct mandatory safety testing of neural networks before their public release.

Previously, his startup had opened its technologies to U.S. intelligence and defense agencies.