Claude Haiku 5.5 launches at about 75% lower running cost than Haiku 4.5

Anthropic released Claude Haiku 5.5 on 7 October. It is the new small, fast Claude model for high-volume work: summaries, classification, database queries, live customer support and browser use. Anthropic says it costs about 75% less to run than Haiku 4.5 on average. For prompts up to 100,000 tokens, input is $0.10 and output is $0.50 per million tokens (Haiku 4.5 was $1 and $5). Above 100,000 tokens it is $0.50 and $2.50. It is the first Haiku with an adjustable effort setting, and it is live now on the Claude API (claude-haiku-5-5), AWS, Google Cloud and Azure. Anthropic's own benchmarks show a big jump over Haiku 4.5, but it still trails Sonnet 5.5 on complex coding, and Anthropic says to keep using the bigger models there. Two more changes came with it. Sonnet 5.5 cache reads are now half price ($0.10 per million tokens), which Anthropic says makes Sonnet about 20% cheaper on most agent work. And this week Max and Team subscribers start getting a monthly API credit: $100 on Max 5x, $200 on Max 20x, and up to $500 pooled for a Team. The Python and TypeScript SDKs also add computer use and browser use in beta. On Hacker News (600+ points), most people welcomed the price, the Sonnet cache cut and above all the API credits. One camp says this finally makes Haiku worth using against OpenAI's cheap GPT-6 Luna, with commenters noting the same price under 100,000 tokens and better scores on Anthropic's charts. Critics call the 100,000-token cutoff too low, point out the new tokenizer uses slightly more tokens per task (Anthropic says so too), and a few hands-on tests found it weak at complex UI design. Some see the credits as a way to pull subscribers into API billing. The rough consensus: a strong, cheap workhorse for subagents, summaries and document lookups, not a replacement for Sonnet or Opus on hard coding. Sorted take: if you run ticket tagging, document Q&A or summaries on a bigger model today, re-test that job on Haiku 5.5 this week and keep prompts under 100,000 tokens to stay on the cheap tier.