Anthropic launches Haiku 4.5, a cheaper and faster AI model for enterprise-scale adoption

Anthropic, the San Francisco-based artificial intelligence company, has unveiled a major update to its smallest AI model, Claude Haiku, as it targets businesses looking for powerful yet cost-effective generative AI tools. The new version, called Haiku 4.5, promises enterprise-grade performance at a fraction of the cost of larger systems.

The company said the updated Haiku model is one-third the cost of its mid-sized model Sonnet 4 and about one-fifteenth the cost of its flagship Opus model, while matching or outperforming Sonnet on several benchmark tasks — including coding and data analysis. The release marks the first significant overhaul of Anthropic’s lightweight model in about a year.

Mike Krieger, Anthropic’s Chief Product Officer, said the company sees growing demand from traditional enterprises outside Silicon Valley that want to integrate AI into their systems but need affordable options.

“Often, there’s a lot of scale to that,” Krieger said. “Small models really help because they can be a more economical way of deploying into that.”

Smaller AI models, such as Haiku, are especially suited for large organizations that deploy AI across hundreds or thousands of internal workflows — where computational costs can multiply quickly. By lowering the cost per query, Anthropic hopes to make its AI accessible to companies that might otherwise hesitate to adopt high-end generative tools.

Anthropic said that around 80% of its revenue now comes from enterprise clients, with more than 300,000 businesses using its tools internally or in customer-facing products. The company’s annual revenue run rate — an estimate based on current sales — has climbed to nearly $7 billion, underscoring the accelerating demand for its Claude family of AI models.

Founded in 2021 by former OpenAI employees, Anthropic has rapidly emerged as one of the most valuable AI startups in the world, with a valuation of about $183 billion. Its Claude models compete directly with OpenAI’s GPT series and Google’s Gemini models in the generative AI market.

The company’s latest strategy reflects a shift across the industry: while early hype centered on ultra-powerful, compute-heavy models costing hundreds of millions of dollars to train, businesses are increasingly turning to smaller, specialized AI systems that can handle practical, day-to-day tasks more efficiently.

According to Krieger, many companies now use a hybrid approach, combining large and small models. “Some rely on the most advanced systems to set strategy or make plans, but outsource the grunt work to smaller models that can search the web and synthesize information,” he said.

With Haiku 4.5, Anthropic aims to strike the balance between capability and cost — appealing to enterprises seeking scalable AI adoption without the heavy compute bills associated with frontier models.