Dario Amodei, co-founder and CEO of Anthropic, at the 2025 World Economic Forum.
Stefan Vermes | Bloomberg | Getty Images
Anthropic on Wednesday announced Claude Haiku 4.5, a compact artificial intelligence model available as a low-cost offering to all of its users.
Anthropic says the model is fast and outperforms other large models that were considered state-of-the-art just a few months ago.
For example, the Claude Haiku 4.5 is a better computer user than the Claude Sonnet 4, a mid-sized model the company launched in May. According to SWE-bench Verified, a test set that measures the software coding capabilities of AI systems, its performance when coding is similar to Claude Sonnet 4 and OpenAI’s latest model GPT-5.
“It punches way above its weight,” Mike Krieger, Anthropic’s chief product officer, said in an interview with CNBC.
Claude Haiku 4.5 is available to Anthropic’s free users, making it the cheapest model available to paid users.
Jack Silva | Null Photo | Getty Images
Anthropic is an AI startup that develops a large family of language models called Claude. The company assigns new numbers to models with each generation, but typically the smallest model in the family is called the Haiku, the mid-sized model is called the Sonnet, and the largest model is called the Opus.
OpenAI emerged in 2022 with the launch of its chatbot ChatGPT, and the following year Anthropic launched a rival product, Claude. It utilizes Anthropic’s family of models, and users can choose between free and paid tiers.
The release of Claude Haiku 4.5 comes just weeks after the company announced Claude Sonnet 4.5 in September and Claude Opus 4.1 in August. Anthropic is working to release another model, likely an updated version of the Opus, by the end of this year or early next year, Krieger said.
For paying users, Haiku models typically cost about one-third of Anthropic’s Sonnet models, while Sonnet models cost one-fifth of Opus models, Krieger said. He added that Anthropic’s free users can still choose to use Claude Sonnet 4.5, but Claude Haiku 4.5 is smaller and offers more space.
Claude Sonnet 4.5 remains Anthropic’s highest-performing model, but the company says Claude Haiku 4.5 is best suited for users looking for fast, accurate answers.
“Even though it’s not as smart as Sonnet, I’ve started using Claude as the default for my own use, especially in mobile apps, because it’s much faster to get answers,” Krieger said.
The two models can also work together. According to Anthropic, Claude Sonnet 4.5 can create multi-step plans to solve complex problems, Claude Haiku 4.5 can complete subtasks within those plans, and more.
Running models in parallel can be especially useful for companies that want to use AI to tackle long-term projects, Krieger said.
“You can also have Haiku monitor the flow of your financial data. Haiku is a smaller, cheaper, faster model, so it can monitor larger amounts of data. You can then pass those early insights to Sonnet for deeper analysis,” he said.
Anthropic ranks No. 4 on CNBC’s 2025 Disruptor 50 list with a valuation of $183 billion and serves more than 300,000 enterprise customers. An Anthropic spokesperson said the company’s annual revenue run rate is approaching $7 billion this month.
The company has been racing to catch up with competitors such as: google And OpenAI’s valuation has ballooned to $500 billion. Following the release of GPT-5 in August, OpenAI closed a multi-billion dollar infrastructure deal and released a short-form video app called Sora.
Due to the breakneck pace of the industry, Anthropic doesn’t have much time to rest easy after launch. While the company was training for Claude Sonnet 4.5, it had already started work on Claude Haiku 4.5.
“We’re really firing on all cylinders,” Krieger said.
WATCH: Anthropic announces overseas hiring push as competition with OpenAI becomes global
