Quick Answer
- Claude Haiku 5.5 is Anthropic’s cheapest and fastest model. It launched on October 7, 2026.
- Price for prompts up to 100,000 tokens: $0.10 input and $0.50 output per million tokens. That is about 90% lower than Haiku 4.5 ($1 and $5).
- Above 100,000 tokens the price is $0.50 and $2.50 per million.
- It has 1 million tokens of context, a new effort setting, and it reads text and images.
- Available on the Claude API, Amazon Web Services, Google Cloud and Microsoft Azure. API name:
claude-haiku-5-5.
What is Claude Haiku 5.5?
Haiku is Anthropic’s small model line. It is built for fast, high-volume work like summaries, sorting, database queries and helper agents. Haiku 5.5 is the new version, and it is the first Haiku with an adjustable effort setting. Medium is the default.
It reads text and images and gives text back. Its knowledge cutoff is June 2026. For comparison, our guide to Claude Sonnet 5.5 covers the larger mid-tier model.
Claude Haiku 5.5 price
| Haiku 5.5 (up to 100K tokens) | Haiku 5.5 (above 100K) | Haiku 4.5 | |
|---|---|---|---|
| Input | $0.10 | $0.50 | $1.00 |
| Output | $0.50 | $2.50 | $5.00 |
| Cache reads | $0.01 | $0.05 | $0.10 |
Prices are per million tokens. Batch processing gets another 50% off.
Is it really 90% cheaper? Only for short prompts. Anthropic’s own estimate for the average workload is about 75% lower, because the new tokenizer counts roughly 30% more tokens for the same text. Anthropic also said most Haiku 4.5 requests fell into the cheaper tier.
Anthropic also cut Sonnet 5.5 cache reads from $0.20 to $0.10 per million tokens.
Claude Haiku 5.5 vs GPT-6 Luna
Both start at the same short-prompt price: $0.10 input and $0.50 output. The difference is where the higher price starts. Haiku’s higher rate begins above 100,000 tokens. GPT-6 Luna’s begins above 272,000 tokens.
On Anthropic’s own tests, Haiku 5.5 scores higher than Luna:
| Test | Haiku 5.5 | GPT-6 Luna | Sonnet 5.5 |
|---|---|---|---|
| OSWorld 2.1 (computer use) | 72.4% | 48.9% | 83.9% |
| Terminal-Bench 4.0 (command line) | 39.2% | 16.4% | 70.6% |
| GDPval-AA v2.1 (office work) | 1,620 | 1,437 | 1,840 |
These are vendor-reported numbers, not independent tests. Sonnet 5.5 beats Haiku on every listed test. For the bigger price picture, see our Claude Opus 5.5 vs GPT-6 Sol and Luna price war article.
What is new in Haiku 5.5?
- Effort controls: you choose how many tokens the model spends. Terminal-Bench at maximum effort is about 39%, but at medium effort it is about 20%.
- 1M context and up to 128K output tokens.
- Computer and browser use in beta through the Python and TypeScript SDKs.
- Stricter cyber safeguards. Penetration testing is blocked. Broader access needs Anthropic’s verification programs.
- Some settings such as custom
temperature,top_portop_kreturn an error.
Who should use it?
- Good fit: high-volume jobs like summaries, tagging, routing, live customer support and fast helper agents.
- Not the best fit: hard coding or deep reasoning. Use Sonnet 5.5 for that.
- Watch out: other small models can be cheaper. Qwen3.7 Flash costs $0.03 input and $0.13 output for prompts up to 32,000 tokens, and Z.ai’s GLM-5.3-Flash scores 1,647 on GDPval-AA, above Haiku’s 1,620, per an outside tracker.
FAQs
How much does Claude Haiku 5.5 cost?
$0.10 per million input tokens and $0.50 per million output tokens for prompts up to 100,000 tokens. Above that, $0.50 and $2.50.
When was Claude Haiku 5.5 released?
October 7, 2026.
Is Haiku 5.5 better than Sonnet 5.5?
No. Sonnet 5.5 scores higher on every test Anthropic listed. Haiku is cheaper and faster.
What is the Claude Haiku 5.5 API name?
claude-haiku-5-5.
Can I use it for free in the Claude app?
The sources we checked cover the API and cloud platforms. They do not say how it appears in the Claude app.
Is it the same price as GPT-6 Luna?
For short prompts, yes. The higher price tier starts at a different point.
