1. News
  2. Releases
  3. Anthropic · Models
ReleaseAnthropic · Models2 min read

Anthropic releases Claude Haiku 4.5, a faster and cheaper small AI model

Anthropic's new small AI model, Claude Haiku 4.5, is said to match an older top coding model at one-third the cost and twice the speed.

By the Dashline deskSource: Mashable
Anthropic releases Claude Haiku 4.5, a faster and cheaper small AI model
Photo: Pauli Rautakorpi, CC BY 2.0

Anthropic has released Claude Haiku 4.5, a small AI model built to be quick and cheap. The company says it matches the coding skill of its earlier Claude Sonnet 4 model, while costing one-third as much and running twice as fast. It was announced on Wednesday.

What it does

Haiku 4.5 is the newest version of Anthropic's compact model. Anthropic describes its performance as near the level of its best models. It is aimed at developers and companies building tools that need fast replies, such as chat assistants, coding helpers and customer service tools.

What it costs and where to find it

Pricing is $1 per million input tokens and $5 per million output tokens. Tokens are small chunks of text. Input is what you send to the AI, and output is what it writes back. The model is available in Anthropic's apps and in Claude Code. Businesses can also reach it through Amazon Bedrock and Google Cloud Vertex AI.

The safety claim

Anthropic says Haiku 4.5 is its safest model yet, based on its own internal tests. It scored lower on misalignment, meaning behavior its makers did not intend, than even Claude Sonnet 4.5, the company's flagship model. It carries an AI Safety Level 2 rating. That is less restrictive than the ASL-3 label given to Anthropic's more advanced models.

The political backdrop

Mashable notes the timing. Earlier this week, Bloomberg reported that David Sacks, the Trump administration's AI czar, accused Anthropic of "regulatory capture" and "fear-mongering." That was over the company's support for California's new AI transparency law. Cofounder Jack Clark pushed back. He said the company agrees with the White House on most issues but still backs responsible federal oversight.

What it means for business owners

If your business uses AI for chat or customer service, a cheaper and faster model could lower your costs. The speed, cost and safety claims all come from Anthropic. A small test on your own tasks would show whether Haiku 4.5 holds up.

  • Check whether your current AI tools offer Haiku 4.5.
  • Compare its answers with your current model on real questions.
  • Compare the price and response time before switching.
The takeaway

Anthropic says Haiku 4.5 is faster and cheaper than before, so businesses using AI chat may want to test it against their current setup.

More articles

You should really read these.

Next step

Let’s talk