Haiku 5.5: Anthropic’s Fastest, Cheapest Claude Yet
Anthropic Launches Haiku 5.5 as Its Cheapest and Fastest Claude Model Yet
Anthropic has released Claude Haiku 5.5, the third model in its Claude 5.5 lineup, expanding its product range as it builds out a broader artificial intelligence portfolio ahead of a planned initial public offering (IPO).
The launch follows recent releases of Claude Opus 5.5 and Claude Sonnet 5.5, completing a three-tier family designed to serve different enterprise and developer workloads. Anthropic said Haiku 5.5 is built for high-volume, cost-sensitive usage, including repetitive, large-scale tasks.
Haiku 5.5 is being positioned as the low-cost option in the Claude 5.5 lineup. Reuters reported that it costs about 75% less than its predecessor, Claude Haiku 4.5, on average.
- Prompts under 100,000 tokens: $0.10 per million input tokens and $0.50 per million output tokens
- Longer prompts: $0.50 per million input tokens and $2.50 per million output tokens
According to Anthropic’s documentation, Haiku 5.5 supports a context window up to one million tokens and can generate outputs of up to 128,000 tokens. Anthropic also described it as its fastest Claude model to date, targeting tasks such as summarization, classification, customer support, browser use, and coding subagents.
Anthropic said Haiku 5.5 is available across major cloud platforms including AWS, Google Cloud, and Microsoft Azure, extending its reach for teams that run production workloads in those ecosystems.
The release underscores a broader trend in enterprise AI: model providers are differentiating product tiers by cost, speed, and context length to meet operational needs. In that framing, Haiku 5.5 is aimed at organizations that prioritize throughput and predictable spend while still requiring strong performance for everyday automation and support workflows.
