Tech Desk 
Oct 16, 2025
Anthropic has launched Claude Haiku 4.5, its most efficient and affordable AI model yet. The new version introduces frontier-level AI capabilities to a wider audience, providing faster responses and lower costs for developers and businesses.
According to Anthropic, Chief Product Officer Mike Krieger, the model enables teams to deploy AI across multiple workflows more easily.
“Often, there’s a lot of scale to that,” he said, describing how the new model allows wider use while keeping budgets in check.
Anthropic Haiku 4.5 model
Anthropic Haiku 4.5 offers almost the same coding effectiveness as Anthropic's higher-tier model Sonnet 4. However, it operates at more than double the speed and roughly one-third the cost.
It is especially well-suited for chat assistants, automation tools, and multi-agent systems, where speed matters the most.
Haiku 4.5 is available now through the Claude API as claude-haiku-4-5 and is priced at $1 per million input tokens and $5 per million output tokens.
Anthropic's objective is to win over startups and businesses looking for economically viable AI at scale without sacrificing performance.
Reportedly, the release can help democratise AI availability. Developers expect to build Haiku 4.5 to support tasks such as customer support bots, thin coding assistants, and automated workflow management systems.
Anthropic is also transparent and clear that Haiku 4.5 will not match larger models like Sonnet 4.5 or Opus in deep reasoning or project-based work.
Nonetheless, Anthropic's objective remains clear: scale AI, while optimising speed, cost, and performance.
