Anthropic Releases Claude Haiku 4.5 With Faster Performance and Lower Costs

Anthropic has released Claude Haiku 4.5, the latest version of its smallest AI model, announced in a company blog post. The model delivers near-frontier coding performance comparable to Claude Sonnet 4, while operating at one-third the cost and more than twice the speed.

According to Anthropic, Claude Haiku 4.5 is available immediately through the Claude API as well as on Amazon Bedrock and Google Cloud’s Vertex AI. Pricing is set at $1 per million input tokens and $5 per million output tokens. The company’s benchmarks show the model achieving 73% on SWE-Bench Verified and 41% on Terminal-Bench, placing it on par with OpenAI’s GPT-5 and Google’s Gemini 2.5 in similar tests.

Claude Haiku 4.5 will serve as the default model for all free Anthropic plans. Its lightweight design allows it to run multiple parallel agents or operate in combination with larger models like Sonnet 4.5, enabling complex orchestration setups. The model is aimed at real-time applications such as chat assistants, customer support, and code generation, where speed and responsiveness are critical.

Anthropic reports that Haiku 4.5 is its safest model to date, earning an AI Safety Level 2 classification after showing lower rates of misaligned behavior than previous releases. The company noted that the model’s efficiency and safety improvements make it suitable for both production environments and large-scale deployment scenarios.