Claude Haiku 5.5 cuts small-model API pricing by up to 90%
Anthropic's new small model adds an adjustable effort setting and a Sonnet-matching tokenizer, with input pricing from $0.10 per million tokens.
Anthropic released Claude Haiku 5.5 on 7 October, calling it the cheapest, fastest and most capable small model it has shipped. It’s live now through the API as claude-haiku-5-5, on AWS, Google Cloud and Microsoft Azure, and on Claude.ai for Free, Pro, Max, Team and Enterprise plans.
Pricing is tiered by prompt length: $0.10 per million input tokens and $0.50 per million output tokens for prompts up to 100,000 tokens, rising to $0.50 and $2.50 beyond that. Cache reads start at $0.01 per million tokens. Against Haiku 4.5, Anthropic says that’s roughly 90% cheaper under the 100k-token threshold and 50% cheaper above it. The model also gets an adjustable effort setting — a first for Haiku-class Claude — and a tokenizer now aligned with Sonnet and Opus. Anthropic’s own benchmarks put it at 72.4% on OSWorld 2.1 and 39.2% on Terminal-Bench 4.0.
Why it matters: a 90% price cut changes the calculus for high-volume work — classification, summarisation, browser and computer-use agents, live customer support — where cost previously forced a choice between Haiku’s speed and a pricier model’s accuracy. At these rates that trade-off mostly disappears.