Anthropic has just dropped Claude Haiku 4.5, and it’s making waves in the AI world. The company calls it its “fastest and most efficient small model yet,” promising Sonnet 4 level performance at one-third the cost — and twice the speed.
According to Anthropic’s latest blog post, the new Haiku 4.5 matches the capabilities of larger models like GPT-5, Gemini 2.5, and Claude Sonnet 4 across a range of benchmarks, while staying lightweight enough for real-time applications.
Benchmarks and performance
Anthropic’s internal tests show Haiku 4.5 scored 73% on SWE-Bench Verified and 41% on Terminal-Bench, placing it on par with Sonnet 4 and other top-tier models. It also delivers strong results in tool use, computer use, and visual reasoning — areas where smaller models often struggle.
Ideal for developers and free AI tools
The company says Claude Haiku 4.5 will be immediately available under all free Anthropic plans, making it perfect for developers and startups looking to add intelligence to apps without huge infrastructure costs.
Because it’s lightweight, it can run multiple AI agents in parallel or work alongside larger models like Sonnet for hybrid AI systems.
Anthropic’s Chief Product Officer Mike Krieger said Haiku opens up new deployment possibilities:
“We’re giving people a complete agent toolbox where each model has the right mix of intelligence, speed, and cost. Haiku-powered sub-agents can now execute at lightning speed while Sonnet handles complex reasoning.”
Real-world use cases
One of the biggest opportunities lies in AI coding tools like Claude Code, where response speed is critical. Early partners — including Zencoder CEO Andrew Filev — say Haiku 4.5 will “unlock an entirely new set of use cases,” especially for real-time development and automation tasks.
A steady stream of innovation
Claude Haiku 4.5 follows closely on the heels of Claude Sonnet 4.5 (released two weeks ago) and Opus 4.1 (launched two months ago). The original Haiku debuted in October 2024, and this upgrade cements Anthropic’s reputation for quickly iterating its AI ecosystem.
With its balance of speed, affordability, and intelligence, Claude Haiku 4.5 could become the go-to model for developers building fast, efficient AI applications — especially in environments where latency and cost matter most.