Anthropic launches Claude Haiku 4.5, a smaller, cheaper AI model
Dario Amodei, co-founder and chief executive officer of Anthropic, at the World Economic Forum in 2025.
Stefan Wermuth | Bloomberg | Getty Images
Anthropic on Wednesday announced Claude Haiku 4.5, a small AI model available as a lower-cost offering to all of the company’s users.
Anthropic says the model is fast and can outperform other larger models that were considered cutting-edge just months ago.
The Claude Haiku 4.5 is better at computing than, for example, the Claude Sonnet 4, a midsize model the company released in May. According to SWE-bench Verified, a test suite that measures the software coding capabilities of an AI system, it performs similarly to Claude Sonnet 4 and OpenAI’s latest model, GPT-5, in coding.
“It punches well above its weight,” Anthropic’s chief product officer Mike Krieger told CNBC in an interview.
Claude Haiku 4.5 is available to Anthropic’s free users and is now the cheapest model available to paid users.
Jaque Silva | Nurfoto | Getty Images
Anthropic is an AI startup that has developed a family of large language models called Claude. The company gives new numbers to models as generations pass, but the smallest model in the family is usually called Haiku, the medium-sized model is called Sonnet, and the largest model is called Opus.
After OpenAI burst onto the scene with the launch of chatbot ChatGPT in 2022, Anthropic launched a competing product, Claude, the following year. It is powered by Anthropic’s family of models, and users can choose between free and paid tiers.
The launch of Claude Haiku 4.5 comes just a few weeks after the company announced Claude Sonnet 4.5 in September and Claude Opus 4.1 in August. Krieger said Anthropic is trying to launch another model, possibly an updated version of the Opus, by the end of this year or early next year.
Krieger said that for paid users, Haiku models are about one-third the cost of Anthropic’s Sonnet models, while Sonnet models are one-fifth the cost of Opus models. Anthropic’s free users can still choose to use Claude Sonnet 4.5, but they will get more capacity from Claude Haiku 4.5 because it is smaller, he added.
Claude Sonnet 4.5 is still Anthropic’s best-performing model, but the company said Claude Haiku 4.5 is ideal for users looking for fast and accurate answers.
“Even for my own use, even though it’s not as smart as Sonnet, I’ve started using it by default in Claude, especially on the mobile app, because it’s a lot faster to get a response,” Krieger said.
The two models can also work together. Anthropic said Claude Sonnet 4.5 can create multi-step plans to solve complex problems, and Claude Haiku 4.5, for example, can complete subtasks within those plans.
Krieger said running models in parallel could be especially useful for businesses looking to use AI to tackle long-term projects.
“You can have Haiku monitor financial data streams — and because it’s a smaller, cheaper, faster model, it can do so at a higher volume — and then feed its initial insights into Sonnet for deeper analysis,” he said.
Anthropic, ranked No. 4 on CNBC’s 2025 Disruptor 50 list and valued at $183 billion, serves more than 300,000 commercial customers. Annual revenue is approaching $7 billion this month, according to an Anthropic spokesperson.
Company races to catch up with rivals Google and OpenAI, whose value has risen to $500 billion. Following the launch of GPT-5 in August, OpenAI signed several multibillion-dollar infrastructure deals and released a short-form video application called Sora.
The breakneck pace of the industry doesn’t give Anthropic much time to adjust after launch. While the company was conducting training on Claude Sonnet 4.5, it had also started work on Claude Haiku 4.5.
“We’re really firing on all cylinders,” Krieger said.
WRISTWATCH: Anthropic announces recruitment move abroad as competition with OpenAI goes global
