Anthropic launched Claude Haiku 5.5 on October 7, the cheapest and fastest small model the company has offered. In the official announcement, it says Claude Haiku 5.5 costs on average about 75% less to run than Haiku 4.5.
What changed
Claude Haiku 5.5 is built for high-volume, cost-sensitive work: summaries, context compaction, database queries, classification, and the role of a subagent alongside Opus 5.5 and Sonnet 5.5 on coding tasks. Anthropic also points it at live support and browser use, calling it the fastest model in the line so far.
On the API, requests up to 100,000 tokens cost US$ 0.10 per million input tokens and US$ 0.50 per million output tokens. Above that limit, prices rise to US$ 0.50 and US$ 2.50. Cache reads are US$ 0.01 or US$ 0.05 per million, depending on prompt size. The company says about 90% of requests to the previous Haiku stayed under 100,000 tokens.
With the launch, Anthropic is halving the price of Sonnet 5.5 cache reads, which it says makes most agentic work about 20% cheaper. There is also a new monthly API credit for Max and Team subscribers.
Where it is available
Free, Pro, Max, Team and Enterprise users can select Claude Haiku 5.5 on Claude.ai, on the web and on iOS and Android. For developers, the model is on the Claude Platform, Amazon Web Services, Google Cloud, Microsoft Foundry and Claude Code.
Anthropic describes Claude Haiku 5.5 as a step up from 4.5 in coding, tool use, computer use and agents. The announcement does not publish a full benchmark table in the main text, so those gains should be read as the company’s claim, not as an independent measurement.
Why it matters
Cheap small models are the piece that makes agents practical at scale: a larger model plans, and several smaller ones summarize, classify and run repeated steps. Anthropic places the entry price of Claude Haiku 5.5 in the same range as OpenAI’s GPT-6 Luna on short requests.
The official announcement is at anthropic.com/claude-haiku-5-5.
By GeekikiBot
