---
title: "Claude Haiku 5.5 in Langdock: EU hosting, price, limits"
description: "How to use Claude Haiku 5.5 from Anthropic in Langdock: hosting region, context window, token prices and when to choose it over other models."
url: "https://langdock.com/models/claude-haiku-5-5"
language: "en"
---

1. [Langdock](https://langdock.com/)
2. [AI models](https://langdock.com/models)
3. Claude Haiku 5.5

# Claude Haiku 5.5 in Langdock

Claude Haiku 5.5 is the fastest model in Anthropic’s latest Claude lineup, built for high-volume, latency-sensitive tasks. In Langdock it is available with hosting in the EU. You can use it in chat, in agents and workflows, and via the API, where it costs €0.10 per 1 million input tokens and €0.50 per 1 million output tokens.

Written by LangdockUpdated November 18, 2026

On this page

What is Claude Haiku 5.5 good at?

What are the key facts for Claude Haiku 5.5 in Langdock?

When should you choose Claude Haiku 5.5, and when another model?

How do you use Claude Haiku 5.5 in Langdock?

Where is Claude Haiku 5.5 hosted, and what happens to your data?

Frequently asked questions

Sources

## What is Claude Haiku 5.5 good at?

Anthropic names these typical uses for Claude Haiku 5.5[1]:

- Classification
- Routing
- Extraction
- Subagent tasks

This is how Anthropic describes the model, not a test result of ours. Which tools the model can use in Langdock depends on the tools and integrations enabled in your workspace.

## What are the key facts for Claude Haiku 5.5 in Langdock?

| Fact | Value |
| --- | --- |
| Provider | Anthropic[3] |
| Name in the model selector | Haiku 5.5[3] |
| Hosting region in Langdock | EU[3] |
| Price per 1M input tokens | €0.10[3] |
| Price per 1M output tokens | €0.50[3] |
| Context window documented by Anthropic | 1M tokens[1] |
| Maximum output documented by Anthropic | 128K tokens[1] |
| Input types documented by Anthropic | Text and images[1] |
| Knowledge cutoff documented by Anthropic | June 2026[1] |

The limits Anthropic documents apply to its own API; Langdock can set its own. The context window indicator next to the model selector shows the window size in use.

Token prices apply to the API, to model use in workflows and to [extra usage](https://docs.langdock.com/en/admin/manage-usage/extra-usage), excluding VAT. They are converted from US dollars at the current exchange rate, so they can change slightly from day to day. In chat and agents, Core and Work seats with included models have no usage-based costs; [pricing](https://langdock.com/pricing) lists the plans.

## When should you choose Claude Haiku 5.5, and when another model?

In its latest lineup of four models, Anthropic lists Claude Haiku 5.5 as the fastest and at the lowest token price. For everyday coding, analysis and writing Anthropic describes Claude Sonnet 5.5 as the best combination of speed and intelligence, and for long-running agentic coding and knowledge work it points to Claude Opus 5.5.[1][2]

The table compares what Langdock lists for Claude Haiku 5.5 and for the alternatives. The [model list](https://langdock.com/models) has every model.

| Model | Provider | Hosting region | Input / output per 1M tokens |
| --- | --- | --- | --- |
| Claude Haiku 5.5 | Anthropic | EU | €0.10 / €0.50 |
| [Claude Sonnet 5.5](https://langdock.com/models/claude-sonnet-5-5) | Anthropic | EU | €2 / €10.02 |
| [Claude Opus 5.5](https://langdock.com/models/claude-opus-5-5) | Anthropic | EU | €4.01 / €20.04 |
| [GPT-6.1 Terra](https://langdock.com/models/gpt-6-1-terra) | OpenAI | EU | €1 / €5.01 |

## How do you use Claude Haiku 5.5 in Langdock?

- **Chat:** choose Haiku 5.5 in the model selector of the chat input. You can switch models in the middle of a conversation and set a personal default model.
- **Agents:** choose the model in the [agent editor](https://docs.langdock.com/en/using-langdock/agents/configuration).
- **API:** the [Completion API](https://docs.langdock.com/en/developer/completion-api/completion-overview) has an endpoint that follows Anthropic’s own API format, with the region in the path. Its models endpoint lists the models available to your workspace in that region. The [Agents API](https://docs.langdock.com/en/developer/agents-api/agent-models) lists the models you can use for agents.
- **Admins:** workspace admins enable or disable each model in the workspace settings and choose the default model for new chats and agents. If you do not see the model, ask your admin.

Details are in the [model guide](https://docs.langdock.com/en/using-langdock/models-and-limits/models) of the documentation.

## Where is Claude Haiku 5.5 hosted, and what happens to your data?

The [Langdock model list](https://langdock.com/models) shows the hosting region EU for Claude Haiku 5.5[3]. In a new workspace only models hosted in the EU are enabled by default; admins decide which other models are switched on[8].

For all models, Langdock states that customer data is never used to train or improve AI models[9], and it publishes a [data processing agreement (DPA)](https://langdock.com/dpa) [10]. A hosting region alone does not make a use case GDPR-compliant: the guide to [GDPR-compliant AI](https://langdock.com/gdpr-compliant-ai) explains what Langdock provides and what remains your responsibility.

## Frequently asked questions

### Is Claude Haiku 5.5 hosted in the EU in Langdock?

Yes. The Langdock model list shows the hosting region EU for Claude Haiku 5.5.

### How much does Claude Haiku 5.5 cost in Langdock?

Via the API, in workflows and as extra usage, Claude Haiku 5.5 costs €0.10 per 1 million input tokens and €0.50 per 1 million output tokens, excluding VAT. In chat and agents, Core and Work seats with included models have no usage-based costs.

### How large is the context window of Claude Haiku 5.5 in Langdock?

Anthropic documents a context window of 1M tokens for its own API. Langdock can set its own limit; the context window indicator next to the model selector shows the size in use.

### Is our data used to train Claude Haiku 5.5?

No. Langdock states that customer data is never used to train or improve AI models.

## Sources

1. [Anthropic: Claude Haiku 5.5 model page](https://platform.claude.com/docs/en/models/haiku-5-5/overview) (checked on November 18, 2026)
2. [Anthropic: Models overview](https://platform.claude.com/docs/en/models/overview) (checked on November 18, 2026)
3. [Langdock: Models with hosting region](https://langdock.com/models)
4. [Langdock documentation: Model guide](https://docs.langdock.com/en/using-langdock/models-and-limits/models)
5. [Langdock documentation: Agent configuration](https://docs.langdock.com/en/using-langdock/agents/configuration)
6. [Langdock documentation: Completion API](https://docs.langdock.com/en/developer/completion-api/completion-overview)
7. [Langdock documentation: Models for the Agents API](https://docs.langdock.com/en/developer/agents-api/agent-models)
8. [Langdock documentation: Workspace settings, models](https://docs.langdock.com/en/admin/workspace/workspace)
9. [Langdock: Security](https://langdock.com/security)
10. [Langdock: Data processing agreement (DPA)](https://langdock.com/dpa)

Claude Haiku 5.5 is a product of Anthropic. Product names are trademarks of their owners. Statements about the model were taken from Anthropic’s documentation on the date shown in the sources.

## Related pages

- [Claude Sonnet 5.5 in LangdockClaude Sonnet 5.5 is the Claude model Anthropic describes as its best combination of speed and intelligence.](https://langdock.com/models/claude-sonnet-5-5)
- [Claude Opus 5.5 in LangdockClaude Opus 5.5 is Anthropic’s model for long-running agentic coding and knowledge work.](https://langdock.com/models/claude-opus-5-5)
- [GPT-6.1 Terra in LangdockGPT-6.1 Terra is OpenAI’s fast, efficient model for visual reasoning, precise computer use and clear, natural writing.](https://langdock.com/models/gpt-6-1-terra)
- [All AI models in LangdockEvery model with hosting region and token prices.](https://langdock.com/models)
- [GDPR-compliant AI for companiesWhat Langdock provides and what stays with you as the controller.](https://langdock.com/gdpr-compliant-ai)
