← sources

Introducing Claude Haiku 5.5: the cheapest, fastest, and most capable small model we’ve ever released

source: reddit (reddit:ClaudeAI)published 2026-10-07 18:01ingested 2026-10-08 02:36

retained text

Claude Haiku 5.5 is the cheapest, fastest, and most capable small model we’ve ever released. On average, it costs around 75% less to run than Claude Haiku 4.5.

Haiku 5.5 is designed for high-volume, cost-sensitive tasks. It reliably handles quick, repetitive work like summaries and classification, and pairs well with Claude Opus 5.5 and Sonnet 5.5 as a sub-agent on coding work. It’s also fast enough for live customer support and browser use.

It’s a significant step up over Haiku 4.5 across coding, computer use, and knowledge work. It’s also our first Haiku model with an adjustable effort setting, so you can decide whether to optimize for cost or intelligence on each task.

Per token, it costs 90% less than Haiku 4.5 on tasks under 100,000 tokens, which make up around 90% of requests to our previous Haiku model. On longer tasks, it’s half the price.

Haiku 5.5 is available now on all platforms, including Amazon Web Services, Google Cloud, and Microsoft Azure.

We’re also halving the price of cache reads on Claude Sonnet 5.5, which makes it around 20% cheaper to run on most long-running work.

Read more: [https://www.anthropic.com/claude-haiku-5-5](https://www.anthropic.com/claude-haiku-5-5)

document fetched 2026-10-08 00:22 · canonical https://reddit.com/r/ClaudeAI/comments/1x03k4j/introducing_claude_haiku_55_the_cheapest_fastest