News · Published October 16, 2026

Claude Haiku 5.5: What Changed and What It Costs Now

Anthropic released Claude Haiku 5.5 on October 7, 2026, calling it the cheapest, fastest, and most capable small model it has released, according to its own announcement published that day. The headline number is the price: Haiku 5.5 runs up to 90% cheaper than Haiku 4.5 for most requests. Here is what actually changed, what it costs in plain numbers, and who should care.

What Anthropic announced, and when

Per Anthropic's own announcement, published October 7, 2026, Claude Haiku 5.5 is built for high-volume, cost-sensitive tasks: summaries, compactions, database queries, and classification requests. Anthropic says it also pairs well with Opus 5.5 and Sonnet 5.5 as a subagent on coding work, and that it is the company's fastest model to date, which it says suits speed-sensitive tasks like live customer support and browser use. The same announcement says Anthropic is updating its Claude Python and TypeScript SDKs to add beta support for computer use and browser use, and describes Haiku 5.5 as especially well-suited to that combination of speed, capability, and price.

The new pricing, in plain numbers

According to Anthropic's announcement, Claude Haiku 5.5 is priced 90% lower than Haiku 4.5 for requests up to 100,000 tokens, which Anthropic says covered about 90% of Haiku 4.5's traffic, and 50% lower for requests above that. Per million tokens, Anthropic lists input at $0.10 up to 100,000 tokens and $0.50 above that, versus $1.00 for Haiku 4.5 and $2.00 for Sonnet 5.5. Output is $0.50 and $2.50, versus $5.00 for Haiku 4.5 and $10.00 for Sonnet 5.5. Cache reads are $0.01 and $0.05, versus $0.10 for both Haiku 4.5 and Sonnet 5.5. Anthropic also notes that Haiku 5.5 uses an updated tokenizer, similar to Sonnet 5.5's and Opus 5.5's, which means it needs slightly more tokens to do the same task, a factor already folded into its blended "roughly 75% cheaper to run" claim.

The same announcement says Anthropic halved Claude Sonnet 5.5's cache-read price on the same day, from $0.20 to $0.10 per million tokens, which it says cuts the cost of most agentic work on Sonnet 5.5 by around 20%. Anthropic adds that Claude Max and Team subscribers are also getting a new monthly API credit for building on the Claude Platform, though the announcement does not state a dollar amount for that credit.

How it compares to Haiku 4.5 and Sonnet 5.5

Anthropic's announcement says Haiku 5.5 is its first Haiku-class model with an adjustable effort setting, letting users trade cost for intelligence the same way Anthropic's larger models already allow. On the benchmarks Anthropic published, Haiku 5.5 scores 1620 on GDPval-AA v2.1 against 735 for Haiku 4.5, 1437 for what Anthropic labels GPT-6 Luna, and 1840 for Sonnet 5.5 shown "for reference." On Terminal-Bench 4.0, Anthropic reports 39.2% for Haiku 5.5 against 0.0% for Haiku 4.5 and 70.6% for Sonnet 5.5; on OSWorld 2.1's offline subset, 72.4% against 15.7% for Haiku 4.5 and 83.9% for Sonnet 5.5. Anthropic's own read of these numbers is that Sonnet 5.5 and Opus 5.5 remain the better choice for complex agentic coding, with Haiku 5.5 aimed at the simpler, high-volume work around them. On safety, Anthropic says Haiku 5.5 shows major alignment improvements over Haiku 4.5, and that its cybersecurity safeguards are more restrictive than Haiku 4.5's but less restrictive than Sonnet 5.5's. If you are trying to match a task to a model rather than compare benchmarks, our guide to choosing the right Claude model walks through that decision plan by plan.

What this doesn't tell you yet

This article is based on Anthropic's own October 7, 2026 announcement and its linked system card and pricing page, not on independent, third-party benchmark testing. Anthropic did not publish a dollar figure for the new Claude Max and Team API credit, and we have not found independent reporting confirming one. Benchmark scores are Anthropic's own, run at its stated "max effort" setting unless noted otherwise, and may not match results you see in daily use.

Questions people are asking about this

What did Anthropic announce on October 7, 2026?

Anthropic released Claude Haiku 5.5, which it calls the cheapest, fastest, and most capable small model it has released, according to Anthropic's own announcement published that day.

How much cheaper is Claude Haiku 5.5 than Haiku 4.5?

Per Anthropic's own announcement, Haiku 5.5 is priced 90% lower than Haiku 4.5 for requests up to 100,000 tokens, and 50% lower for requests over that, which together average out to roughly 75% cheaper to run.

What is Claude Haiku 5.5 actually built for?

According to Anthropic, it targets high-volume, cost-sensitive work such as summaries, compactions, database queries, and classification, and it pairs with Opus 5.5 or Sonnet 5.5 as a subagent on coding tasks.

Did Claude Sonnet 5.5 also get cheaper on October 7, 2026?

Yes. Anthropic's announcement says it halved Sonnet 5.5's cache-read price from $0.20 to $0.10 per million tokens the same day, cutting the cost of most agentic work on Sonnet 5.5 by around 20%.

Where can I actually use Claude Haiku 5.5?

Anthropic's announcement says it is available now on all platforms, including Amazon Web Services, Google Cloud, and Microsoft Azure, and that developers on the Claude Platform can get started with the model ID claude-haiku-5-5.

Keep reading

Disclosure: this page is an independent reader resource. Some links are Amazon Associates links, which means we may earn a commission from qualifying purchases at no extra cost to you.