Anthropic has launched its latest model, Claude 3.5 Haiku, which is now available to general users through the Claude chatbot on both web and mobile applications as of December 13, 2024.

This new model has achieved a notable 40.6% score on the Software Engineering benchmark, outperforming its predecessor, the 3.5 Sonnet, and OpenAI's GPT-4o.

Introduced alongside the 3.5 Sonnet in October 2024, Claude 3.5 Haiku emphasizes improvements such as lower latency and enhanced instruction-following capabilities.

Notable enhancements include optimized tool usage and reduced latency, catering to both individual users and businesses.

Claude 3.5 Haiku features a significant 200,000-token context window, allowing it to manage extensive inputs better than OpenAI's GPT-4, which has a limit of 128,000 tokens.

Benchmarking indicates that the model has an average time-to-first-token of 0.80 seconds, although it generates tokens at a speed of 65.1 tokens per second, which is slower than some competitors.

Practical applications for Claude 3.5 Haiku include coding recommendations, data extraction, labeling, and content moderation, making it particularly effective for enterprises.

The model excels in user-centric applications and personalized experiences derived from large datasets, enhancing its appeal for enterprise-level solutions.

Claude 3.5 Haiku is accessible through Anthropic's API, Amazon Bedrock, and Google Cloud's Vertex AI, providing flexibility for organizations.

Users can access the model for free, but usage is limited to approximately 10 exchanges per day, depending on server demand.

The rollout of Claude 3.5 Haiku follows significant updates from competitors OpenAI and Google, who have also released new models, intensifying the competitive landscape.

Despite initial expectations for similar pricing to Claude 3 Haiku, the API costs were revised upward due to the enhanced intelligence and performance of the new model.