Anthropic hikes the price of its Haiku model

Anthropic has recently introduced its latest AI model, Claude 3.5 Haiku, which comes with significant enhancements but also a higher price point compared to its predecessor, Claude 3 Haiku. While Claude 3.5 Haiku has shown to match or surpass the performance of its former flagship model, Claude 3 Opus, on various benchmarks, it currently lacks the ability to analyze images.

The unveiling of Claude 3.5 Haiku follows a series of updates from Anthropic, which highlighted the model’s advanced capabilities, applicable to coding suggestions, chatbots, data extraction and labeling, as well as content moderation. However, during the final testing phase, it became apparent that Claude 3.5 Haiku outperformed Claude 3 Opus on several key benchmarks, prompting Anthropic to adjust its pricing strategy to reflect the model’s increased intelligence and capabilities.

The new pricing structure for Claude 3.5 Haiku is set at $1 per million input tokens (approximately 750,000 words) and $5 per million output tokens. In comparison, the pricing for the previous model, Claude 3 Haiku, was significantly lower at 25 cents per million input tokens and $1.25 per million output tokens. This shift represents a notable increase in costs for users.

Moreover, Anthropic’s head of developer relations, Alex Albert, clarified that while Claude 3.5 Haiku is positioned as a more advanced model, it does not include image processing features at launch. As a result, the earlier Claude 3 Haiku model will continue to be available for those who prioritize cost-efficiency and require image processing capabilities.

In summary, Anthropic’s Claude 3.5 Haiku signifies a step forward in AI advances with its enhanced performance, but the substantial price increase and the omission of image analysis at launch may lead some users to reconsider their options.