Claude Haiku 5.5 Launch: Anthropic's Cheapest Model Gets Smarter

Anthropic released Claude Haiku 5.5 on 7 October 2026, describing it as “the cheapest, fastest, and most capable small model we’ve ever released.” It costs $0.10 per million input tokens and $0.50 per million output tokens for prompts up to 100,000 tokens, a tenth of Haiku 4.5’s list price, and it is the first Haiku model with an adjustable effort setting. If you run high-volume classification, extraction, routing or subagent work on the Claude API, this is a release worth testing this week. Source: Anthropic’s Haiku 5.5 announcement and the model documentation.

What Anthropic launched

Haiku 5.5 completes the 5.5 generation, following Opus 5.5 on 22 September and Sonnet 5.5 on 28 September. It is aimed at what Anthropic calls “high-volume, cost-sensitive tasks”: summaries, compaction, database queries and classification, plus speed-sensitive work such as live customer support and browser use. Anthropic also pitches it as a coding subagent working alongside Opus 5.5 or Sonnet 5.5, which handle the planning while Haiku does the many small steps.

The headline specifications from the documentation:

• Context window: 1 million tokens; maximum output 128,000 tokens (up to 300,000 on the Batch API with a beta header).
• Inputs: text and images; output is text.
• Effort levels: Low, Medium, High, Xhigh and Max, with Medium as the API default. Thinking is adaptive and on by default.
• Model IDs: claude-haiku-5-5 on the Claude API, Google Cloud and Microsoft Foundry; anthropic.claude-haiku-5-5 on Amazon Bedrock.
• Retirement: not sooner than 7 October 2027.

Pricing

Per million tokens	Haiku 5.5 (prompts ≤100K / >100K)	Haiku 4.5	Sonnet 5.5
Input	$0.10 / $0.50	$1.00	$2.00
Output	$0.50 / $2.50	$5.00	$10.00
Cache reads	$0.01 / $0.05	$0.10	$0.10
Cache writes (5 min)	$0.125 / $0.625	$1.25	$2.50

Prices are list prices at the time of writing, from Anthropic’s announcement. The Batch API takes another 50% off. Anthropic says Haiku 5.5 costs “around 75% less to run” than Haiku 4.5 on average: about 90% less for prompts up to 100,000 tokens and 50% less above that.

There is one caveat the announcement plays down. The documentation notes that Haiku 5.5 uses a newer tokenizer, so the same text counts as roughly 30% more tokens than on Haiku 4.5. The per-token saving is real, but your per-request saving will be smaller than the per-token table suggests. Run a sample of real traffic through both models before you rewrite your budget.

The API also drops sampling controls: temperature, top_p and top_k should be omitted, and some values return an error. Check the migration guide before switching a production workload.

How it performs, according to Anthropic

Anthropic’s own benchmarks show a large jump over Haiku 4.5, for example 72.4% versus 15.7% on its OSWorld 2.1 subset for computer use, and 39.2% versus 0% on Terminal-Bench 4.0. These are vendor-reported numbers on vendor-chosen tests; treat them as a reason to try the model, not as proof it suits your workload. Our guide to reading AI model announcements explains how to discount launch-day benchmark charts. Anthropic is also clear that Sonnet 5.5 and Opus 5.5 remain stronger for complex agentic coding.

Also announced: cheaper Sonnet caching and API credits

Two changes landed alongside the model:

• Sonnet 5.5 cache reads halved from $0.20 to $0.10 per million tokens, effective 7 October. Anthropic estimates this makes most agentic tasks on Sonnet 5.5 about 20% cheaper.
• Monthly API credits for subscribers. Max 5x subscribers get $100 a month, Max 20x $200, and Team plans up to $500 pooled across users, usable on any model through the Claude Platform. Credits are claimed by linking a Claude Console organisation in billing settings and were rolling out over the week of the launch.

Who should care

Developers paying for high-volume Haiku 4.5 or small-model traffic elsewhere should test Haiku 5.5 now; the price cut is large enough to change architecture decisions, such as running a cheap first pass before escalating to Sonnet. Max and Team subscribers who build small tools should claim the new API credits. For everyday chat users the impact is smaller: Anthropic has not said which Claude app plans offer Haiku 5.5. Our Claude review covers what each plan includes, and our ChatGPT vs Claude comparison helps if you are still choosing an assistant. More launches are tracked in our news section.