Claude 3 Haiku is a large language model developed by Anthropic and released in March 2024 as part of the Claude 3 family, alongside Claude 3 Sonnet and Claude 3 Opus. It is designed to be the fastest and most cost-efficient model in the series, optimized for high-throughput applications.
Like the other Claude 3 models, Haiku is multimodal, able to process both text and image inputs while generating text outputs. It supports a context window of up to 200,000 tokens, with Anthropic noting that the Claude 3 models are technically capable of handling inputs exceeding one million tokens in special cases.
Haiku is positioned as a model well-suited for scenarios that demand speed and scalability at lower cost, such as customer support, summarization, and other tasks where rapid responses are prioritized. Compared to the larger Claude 3 Sonnet and Opus, Haiku provides lower latency and higher efficiency, while the larger models offer stronger reasoning and depth of analysis.